University of Southern California Dissertations and Theses

The Effect Of Explicit Instruction On The Summarization Strategies Of "Underprepared" Native Spanish-Speaking Freshmen In University-Level Adjunct Courses

(USC Thesis Other)

The Effect Of Explicit Instruction On The Summarization Strategies Of "Underprepared" Native Spanish-Speaking Freshmen In University-Level Adjunct Courses

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may be
from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleed through, substandard margins,
and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted. Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in reduced
form at the back of the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6" x 9” black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly to
order.
UMI
A Bell & Howell Information Company
300 North Zceb Road, Ann A rbor MI 48106-1346 USA
313/761*4700 800/521*0600
THE EFFECT OF EXPLICIT INSTRUCTION ON THE
SUMMARIZATION STRATEGIES OF "UNDERPREPARED" NATIVE
SPANISH-SPEAKING
FRESHMEN IN UNIVERSITY-LEVEL ADJUNCT COURSES
by
Lla Diana Kamhi-Stein
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Education)
August 1995
Copyright 1995 Lla Diana Kamhi-Stein
UMI Number: 9614032
Copyright 1995 by
Kamhi-Stein, Lia Diana
All rights reserved.
UMI Microform 9614032
Copyright 1996, by UMI Company. All rights reserved.
This microform edition is protected against unauthorized
copying under Title 17, United States Code.
UMI
300 North Zeeb Road
Ann Arbor, MI 48103
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL ,
UNIVERSITY PARK
LOS ANCELES. CALIFORNIA 9000 ? _ ^
"■/‘•V.
CJ
This dissertation, written by
&ift..D^na..KamhA-Ste.i;n................
under the direction of hex. Dissertation
Committee, and approved by all its members,
has been presented to and accepted by The
Graduate School, in partial fulfillment of re-
quirements for the degree of
DOCTOR OF PHILOSOPHY
Dean of Graduate Sixties
Date
TATION COMMITTEE
lf]l(LL^uJ^uCL (l\A *
Chairperson
ACKNOWLEDGMENTS
j.
This is the end of a Ion? journey that began in
Argentina when my mother enrolled me in EFL classes when I
was eight years old.
There are many people whom I would like to thank for
helping me to reach the end of this journey:
Dr. Behjat Sharif, Dr. Geri-Ann Galanti, and Dr.
Beverly Krilowicz, who allowed me to use their classes for
my research. Special thanks to Dr. Nadine Koch, whose
Political Science class helped me to conceptualize what
later became the topic of my dissertation.
Ann Dywer, who assisted in the data coding, and Connie
Thieraux, who helped me in the data collection and became a
supportive friend in the process.
Dr. Janet Tricamo, Project LEAP co-director and former
director of the Learning Resources Center at California
State University, Los Angeles, for her continuous support of
my study. Linda Zepeda and her student assistants, Irma de
Lira and Irma Ramlrez-Zepeda, who spent many hours obtaining
the records of all the subjects who participated in this
study.
The Office of Academic Technology Support and the
Office of the Dean of the School of Business at California
State University, Los Angeles for their computing resources.
All the students enrolled in the three classes in which
1 collected my data and especially Project LEAP students. I
am proud to have been their teacher*
Former professors who have had considerable influence
on my professional development. Special thanks to Kathi
Bailey, Jose L. GalvAn, Dennis Hocevar, Reynaldo Macias, Pat
Richard-Amato, and Kelly Stromquist.
My husband Alan, for reading the many drafts of this
dissertation, for assisting me to collect my data, for being
the best librarian-husband possible, and for putting up with
me for the last two years. Without Alan, literally, I would
not be here today.
My committee. I could not have asked for a better onel
Thanks to Dr. Robert Rueda for challenging me with questions
that helped me to design a better study. I am grateful to
Dr. Robert Kaplan for his professional and personal support.
He helped me to conceptualize the study, listened to me
patiently, and critiqued many versions of this dissertation.
He is a role model difficult to emulate. I especially wish
to thank my committee chair, Dr. David Eskey for helping me
to keep my b&lance and focus during those times when my
research compass lost its way. Dr. Eskey's guidance was
essential to the successful completion of this dissertation.
Finally, 1 would like to thank Dr. Marguerite Ann Snow. I
am deeply indebted to her for the Project LEAP resources
i ii
that she made available to me. Ann taught me about language
and content, about teaching and learning, about being
generous, about friendship and sharing, about being a
professional, about being a balanced person. I am proud to
have her as my mentor. In the years to come, I hope that my
teaching career will be an ongoing tribute to her
mentorship.
I wish to conclude this journey with this thought:
tMami, pensar que vos s61o queries gue el inglds me sirviera
como un anna laboral en Argentina!
TABLE OP CONTENTS
ACKNOWLEDGMENTS........................................ ii
LIST OF TABLES..........................................ix
LIST OF FIGURES....................................... xii
ABSTRACT..............................................xiii
CHAPTER I: THE PROBLEM
Introduction....................................... l
Background of the Problem...........................2
The Academic Literacy Demands of the
General Education Curriculum....................2
contextualizing Summary Writing in
General Education Courses........................3
The Issue of the Language Readiness of
Non-Native Speakers of English in the
California State University System...........4
The Response of One CSU campus: A
Theoretically-Grounded Approach to the
Instruction of "Underprepared"
Non-Native English Speakers.................. 6
Statement of the Problem..........................10
Purpose of the Study..............................13
Importance of the Study...........................14
Research Questions................................ 16
Overview of the Methodology of this Study.........17
Limitations....................................... 20
Definition of Terms...............................21
Organization of the Remainder of this
Dissertation................................. 23
CHAPTER II: REVIEW OF PERTINENT LITERATURE
Reading-to-write As a Constructive Act........... 25
Summary Writing................. 29
Cognitive Operations............................29
Classification of Summaries.....................33
Empirical Research in summary Writing..........36
Second Language Research and Summary
Writing...................................... 40
Text Structure.................................... 46
Empirical Research on the Effect of
Text Structure...............................47
Linguistic Cohesion in Text.......................49
Explicit Instruction..............................51
Content-Based Second Language Instruction........53
The Adjunct Model: Empirical Research......56
Previous Studies Related to the Student
Population Represented in this Study........60
Summary of this Chapter........................... 65
Notes...............................................65
CHAPTER III: METHODOLOGY
Subjects. ..............................68
Selection of Subjects........................... 72
Equivalence of the Three Groups in
Content Knowledge............................ 75
Equivalence of the Three Groups in
Awareness of the Task Demands................76
The Intervention...................................77
Procedures....................................... 77
Materials for Summarization-Strategy
Training...................................... 80
Instrumentation....................................80
Materials........................................ 80
Data Collection Procedures........................ 82
The Reading Task.................................82
The Summarization Task.......................... 82
Data Analyses: Discourse Analytic Procedures....84
Identification of Idea Units and Main Ideas....84
Holistic Assessment..............................87
Summarization Efficiency Assessment............ 89
Analytic Assessment............................. 92
Statistical Analyses...............................97
CHAPTER IV: FINDINGS
Holistic Assessment................................99
Analysis of Main Ideas........................... 102
Summarization Efficiency Assessment............. 106
Writing Strategies Analyses......................108
Health Science 150............................. 108
Cultural Anthropology 250......................121
Animal Biology 155............................. 130
Summary of Findings.........'..................... 144
CHAPTER V: DISCUSSION OF FINDINGS
Holistic Performance..............................147
Main Ideas and Summarization Efficiency......... 157
Writing Strategies................................165
Combination Strategies......................... 166
Reproduction Strategies....................... 172
Explicit Instruction.............................177
Summary of Discussion of Findings...............179
CHAPTER VI: SUMMARY OF THE STUDY, CONCLUSIONS, AND
RECOMMENDATIONS
Summary of the Study............................ 183
Purpose of the Study....................... 183
Summary of Findings........................... 184
Conclusions......................................190
Recommendations..................................195
Recommendations for Classroom Pedagogy........195
Recommendations for Further Research..........198
Notes............................................ 201
BIBLIOGRAPHY..........................................202
APPENDIXES
A. Student Profile Questionnaire....................227
B. Equivalence in Content Knowledge:
Question for Health Science 150............231
Question for cultural Anthropology 250.....232
Question for Animal Biology 155............233
C. Summary Writing, Awareness of Task Demands...... 234
D. Handout: Strategies for Summary Writing........235
E. Handout: Eight Steps to Evaluating a
Summary..........................................236
F. Handout: Connecting Words............ 237
G. Articles:
The Safer Sex: Health Science 150.........238
Murder in Good Company: Cultural
Anthropology 250...................... 242
Here Comes the Sun: Animal Biology 155....246
H. Instructions for the Reading Task................250
I. Summarization Prompt:
Health Science 150......................... 251
cultural Anthropology 250.................. 252
Animal Biology 155......................... 253
vii
J. Propositions Used by the Expert Readers
to Determine the Level of Importance
of the Ideas:
The Safer sex...............................254
Murder in Good Company..................... 261
Here Comes the Sun......................... 269
K. Main Ideas Identified in the Articles:
The Safer Sex...............................277
Murder in Good Company....... 279
Here Comes the Sun......................... 280
L. Idea Units Identified in the Articles:
The Safer Sex...............................281
Murder in Good Company..................... 282
Here Comes the Sun......................... 283
M. Holistic Rating Scale for Summary Writing....... 284
viii
LIST OF TABLES
Table 1 Denographic Information on the
Subjects...... _ .............................. 70
Table 2 Subjects' Perceptions of Their
Proficiency in English.......................71
Table 3 Subjects' Perceptions of Their
Proficiency in Spanish.......................72
Table 4 Hone Language Background.....................75
Table 5 Means and Standard Deviations for
the Holistic Scores.........................100
Table 6 Analysis of Covariance for the
Holistic Scores............................. 101
Table 7 Means and Standard Deviations for
Main Ideas by Group.........................102
Table 8 Total Number and Percentage of Main
Ideas Reproduced by Group....... 103
Table 9 Analysis of Covariance for the Number
of Main Ideas Reproduced....................105
Table 10 Efficiency of Summarization Mean
Ratio: Number of Main Ideas to the
Number of Words By Group....................106
Table 11 Main Ideas in Health Science:
Total Number and Percentage by Group.......109
Table 12 Adjacent Main Ideas in Health Science:
Total Number and Percentage by Group.......110
Table 13 Conjunctive Relations between Two
Adjacent Main Ideas in Health Science:
Frequency and Percentage by Group..........Ill
Table 14 Means, Standard Deviations, and Total
Number of Idea Units Contained in the
Main Ideas Reproduced in Health Science....ill
Table 15 Main Idea Efficiency in Health Science:
Mean Ratios, Number of Main Ideas to
the Number of Idea Units by Group..........112
ix
Table 16 Individual Idea Units in Health Science:
Frequency and Percentage by Group..........114
Table 17 Idea Unit Combinations within
Paragraphs in Health Science:
Frequency and Percentage by Group 116
Table 18 Idea Unit Combinations across
Paragraphs in Health Science:
Frequency and Percentage by Group..........118
Table 19 Idea Unit Reproduction Strategies
in Health Science: Frequency
and Percentage by Group.....................120
Table 20 Main Ideas in Cultural Anthropology:
Total Number and Percentage by Group....... 122
Table 21 Means, Standard Deviations, and
Total Number of Idea Units
Contained in the Main Ideas Reproduced
in Cultural Anthropology....................123
Table 22 Main Idea Efficiency in Cultural
Anthropology: Mean Ratios,
Number of Main Ideas to the Number
of Idea Units by Group......................124
Table 23 Idea Unit Combinations within
Paragraphs in Cultural Anthropology:
Frequency and Percentage by Group..........127
Table 24 Idea Unit Reproduction Strategies
in Cultural Anthropology:
Frequency and Percentage by Group..........129
Table 25 Main Ideas in Animal Biology:
Total Number and Percentage by Group.......131
Table 26 Adjacent Main Ideas in Animal Biology:
Total Number and Percentage by Group.......132
Table 27 Conjunctive Relations between Two
Adjacent Main Ideas in Animal Biology:
Frequency and Percentage by Group..........133
Table 28 Means, Standard Deviations, and Total
Number of Idea Units Contained in the
Main Ideas Reproduced in Animal Biology....134
x
Table 29 Main Idea Efficiency in Animal Biology:
Mean Ratios, Number of Main Ideas to the
Number of Idea Units by Group...............135
Table 30 Individual Idea Units in Animal
Biology: Frequency and Percentage
by Group..................................... 136
Table 31 Idea Unit Combinations within
Paragraphs in Animal Biology:
Frequency and Percentage by Group.......... 138
Table 32 Idea Unit combinations across
Paragraphs in Animal Biology:
Frequency and Percentage by Group...........141
Table 33 Idea Unit Reproduction Strategies
in Animal Biology: Frequency
and Percentage by Group..................... 143
xi
LIST OF FIGURES
FIGURE 1. Structure of the Writing Model............. 26
a .
FIGURE 2. Health Science: Frequencies of
idea unit combinations within
paragraphs by group........................ 115
FIGURE 3. Health Science: Frequencies of
idea unit combinations across
paragraphs by group........................ 118
FIGURE 4. Health Science: Frequencies of
idea unit reproduction strategies
by group................................... 119
FIGURE 5. Cultural Anthropology: Frequencies of
idea unit combinations within
paragraphs by group........................ 126
FIGURE 6. Cultural Anthropology: Frequencies of
idea unit reproduction strategies
by group................................... 128
FIGURE 7. Animal Biology: Frequencies of
idea unit combinations within
paragraphs by group........................ 137
FIGURE 8. Animal Biology: Frequencies of
combinations across
paragraphs by group........................ 140
FIGURE 9. Animal Biology: Frequencies of
idea unit reproduction strategies
by group................................... 142
xii
ABSTRACT
This study examined whether "underprepared" non-native
speakers of English, university-level native Spanish
speaking freshmen, would benefit from explicit, content-
based instruction in summary writing in adjunct courses.
Pre- and post-intervention summaries written by
"underprepared" native Spanish-speaking freshmen in an
Experimental Group and subjects in two Control Groups— one
consisting of "underprepared" freshmen and the other
consisting of "able" readers— were analyzed for (1) holistic
performance, (2) the number of main ideas included in the
summaries, and (3) the strategies used to reproduce and to
combine the main ideas in the source text.
An Analysis of Covariance (ANCOVA) revealed that the
summaries produced by the subjects in the Experimental Group
achieved significantly higher holistic scores than those
produced by the "underprepared" subjects in Control Group 1
and than the "able" subjects in Control Group 2. A second
ANCOVA test showed that the summaries produced by the
Experimental subjects included a significantly higher number
of main ideas than those written by the "underprepared"
subjects in Control Group 1.
Following instruction, the Experimental subjects did
not gain in the production of abstraction operations that
would help them to understand the global meaning of the
xiii
text; instead, the Experimental subjects read at the
paragraph level. The Experimental subjects, however,
paraphrased more often and copied and nearly copied less
frequently than the "underprepared" subjects in Control
Group 1 and the "able" subjects in Control Group 2.
The overall results of this investigation indicate that
"underprepared" students from a Spanish-speaking background
can benefit from explicit summarization instruction grounded
in the content areas.
CHAPTER I
THE PROBLEM
Introduction
Increasing numbers of non-native English-speaking
freshmen who do not possess college-level English
proficiency are entering U.S. colleges and universities.
These students must perform in a competitive academic
environment that requires them to demonstrate high levels of
academic literacy skills in English. This study
investigated the extent to which cognitive strategy
instruction in an adjunct, second language program helped
"underprepared" freshmen improve their academic skills in
English.
Chapter I of this dissertation provides a description
of the academic literacy demands faced by undergraduate
students and provides an explanation of how summary writing
is used in the general education curriculum. The chapter
presents demographic information about the "underprepared"
language minority student population within the California
State University (CSU) system in general, and within one
campus, California State University, Los Angeles (CSLA) in
particular. Chapter I also provides the statement of the
problem and the importance and the purpose of the study,
followed by specific research questions. In addition,
Chapter I presents an overview of the methodology of this
study, the limitations of this investigation, as well as a
glossary of terms used in this study. Chapter I concludes
with a brief explanation of the organization of the
remainder of this dissertation.
Background of the Problem
The Academic Literacy DemandB_of_the General Education
Curriculum
The literacy demands of the undergraduate curriculum
are extensive. Students in general education courses must
read large amounts of cognitively and linguistically
demanding material. They also have to take notes and
integrate new information with previously read material.
Furthermore, the undergraduate curriculum requires students
to fulfill writing assignments including book and article
reviews, reactions to readings, in-class essays and take-
home examinations, short and long research papers and lab
reports, case studies, and papers requiring the ability to
interpret and to connect theory and data (Braine, 1995;
Bridgeman & Carlson, 1984; Carson, Chase, Gibson, &
Hargrove, 1992; Horowitz, 1986; Ostler, 1980; Rusikoff,
1994; Shih, 1992; The Offices of Public Affairs & Analytical
studies, CSLA, 1994).
Although writing tasks may differ from discipline to
discipline and even from one content professor to the next,
2
very seldom are students required to write essays based on
personal experience (Carson et al., 1992; Rusikoff, 1994;
shih, 1986). In fact, most writing assignments, and
certainly all of the tasks listed above, involve the use of
reading materials as a basis for producing written text.
Contextualizing Summary Writing in General Education courses
Critical to all reading-to-write tasks is
summarization, requiring the comprehension of written text
and the subsequent condensation and transformation of
important information at the expense of unimportant ideas
(Ackerman, 1989; Braine, 1995; Brown & Day, 1983; Brown,
Day, & Jones, 1983; Charry, 1987; Johns, 1988a; Johnson,
1983; Kaplan, Cantor, Hagstrom, Kamhi-Stein, Shiotani, &
Zimmerman, 1994; Kintsch & van Dijk, 1978; Mushakoji, 1991;
Rinaudo, 1993; Sherrard, 1989; Solarz, 1994; Wilhelm, 1993).
While summarization is a cognitively-demanding activity
for first language (Ll) readers and writers, the cognitive
load is even greater for second language (L2) students.
Second language research has shown that L2 proficiency,
content knowledge, cultural and formal schemata, audience
awareness, and metacognitive and cognitive skills affect the
successful completion of reading-to-write tasks, including
summary writing (Basham, 1986; Carrell, I984ab; Carre11 &
Eisterhold, 1983; Johns, 1985; Johns & Mayes, 1990; Kirkland
3
& Saunders, 1991; Kozminsky & Graetz, 1986; Sarig, 1987;
Sherrard, 1989)* However, it is often assumed that, by the
time students enter college, summarization strategies have
already been acquired, and subsequently, students seldom
receive formal instruction in summary writing.
The issue of the Language Readiness of Non-Hative Speakers
of English in the California state University system
In the California State University (CSU) system, much
as in other urban universities in the United states, the
number of students for whom English is not the LI is growing
rapidly. The CSU Committee on Educational Policy (1995)
noted that, in the Fall of 1993, nearly one in four CSU
first-time freshmen reported that English was their second
language. Of these 5,000 to 6,000 students, only one in six
evidenced college-level English proficiency, as indicated by
the numbers of students required to enroll in developmental
reading and writing courses.
The performance of Latino, including native and non
native Spanish-speaking freshmen, entering college-level
English courses in the CSU system is equally discouraging.
Enrollment statistics for the Fall quarter of 1993 revealed
that, among the CSU freshman student population, 60% of the
Mexican-American students and 52% of other Latinos were
required to enroll in pre-baccalaureate English courses.
4
These non-native speakers of English (NNS), whom the
report refers to as "developmental," may have only recently
exited high school classes for limited English proficient
(LEP) students or may have had limited high school exposure
to college preparatory programs. Therefore, they require
instruction in English language skills needed to succeed not
only in English courses but also in a general education
curriculum. As indicated by the Committee, "developmental"
students differ from "remedial" students in that the latter
have had previous instruction but are in need of additional
work in areas such as English or Mathematics.
"Developmental" students, that is NNS who, based on
their scores on the CSU English Placement Test (EPT), do not
possess the level of academic English proficiency necessary
to succeed in college-level courses, are required to enroll
in developmental English courses. However, there is
disparity among the different CSU campuses as to how
"developmental" (and "remedial") courses are funded,
operated, and taught (Academic Program Improvement Workgroup
on Support for Underprepared Students, 1994).
At the same time, "developmental" courses and, by
extension, "developmental" students, are affected by three
problems which are directly related to this study. The
first problem is what appears to be some reluctance or
ambivalence on the part of the academic departments to
5
accept responsibility for developmental efforts (Academic
Program Improvement Workgroup on Support for Underprepared
Students, 1994). This problem may result in the second
problem, involving the isolation of "developmental"
instructors from those in academic departments. The last
problem that affects "developmental" English programs for
NNS is that the CSU system does not have a consistent policy
regarding the instruction of "developmental" NNS. For
example, while some CSU campuses separate "developmental"
from "remedial" students, others do not (English as a Second
Language Workgroup, 1988); and classroom instructors may not
be prepared to teach NNS who have unique L2 demands.
As will be shown below, one CSU campus— -Los Angeles—
implemented a project that was designed to address— at least
in a small way— -this situation. Although the project
offered some remedies, major solutions remain elusive and
require the enhanced commitment of human and fiscal
resources.
The Response of One CSU Campus; A Theoretically-Grounded
Approach to the Instruction of "Underprepared" Non-Native
English Speakers
Of all CSU campuses, California State University, Los
Angeles is the most ethnically and linguistically diverse.
Of the total student population of nearly 18,000 students,
6
Latinos account for nearly 41%; 28.1% are Asian-
Aroerican/Pacific Islander, 21.2% are White, 9.7% are
African-American, and .5% are Native-American. In addition,
69% of the entering freshmen at CSLA report having a
language other than English as their first language. More
importantly, the CSLA statistics for 1993 indicate that
approximately 82% of the NNS enrolled in freshman courses
required to enroll in pre-baccalaureate reading/writing
courses, based on the students' CSU English Placement Test
(EPT) scores.
The lack of academic literacy skills in English
frustrates mainstream faculty, who often are ill prepared to
teach a linguistically diverse student population (Snow,
1994; Snow & Tricamo, 1991). This problem frequently
requires the Learning Resource Center to assume
responsibility for the instruction of "developmental" NNS,
since the English Department at CSLA does not offer courses
for this student population. Snow and Tricamo (1991) point
out that the lack of language development programs at CSLA
was noted in the Western Association of Schools and Colleges
(WASC) Accreditation Report (1990). According to the WASC
Report, the low success rate on the Writing Proficiency
Examination (WPE)— the CSLA's version of the CSU-mandated
Graduation Writing Assessment Requirement (GWAR)— can be
attributed to the fact that the school is not providing
7
students with language instruction that would help remediate
the students' entering disabilities.
One response to the WASC Report was the design of a
project which received funding from the Fund for the
Improvement of Postsecondary Education (F1PSE). This
project, entitled Project LEAP: Learning**English-for-
Academic-Purposes, incorporates supplemental language
development instruction into selected general education
courses (Snow & Tricamo, 1991). Project LEAP includes four
major components: (1) the Faculty Development component,
(2) the Curriculum Modification component, (3) the
Dissemination and Project Continuity component, and (4) the
Language Development Adjunct Study Group Course component.
The Faculty Development component involves the training of
mainstream faculty in the integration of language
instruction into the general education curriculum. The
Curriculum Modification component involves
institutionalizing curriculum changes within the academic
departments through the action of Educational Policy
Committees. The third component, the Dissemination and
Project Continuity component, involves the dissemination of
project activities, including the development of training
manuals and the production of training videotapes. The
fourth component, the most important one for the purposes of
this investigation, involves the implementation of an
8
adjunct study group course for NNS offered in conjunction
with selected general education courses.
The adjunct study group course draws on two rationales.
First, support for the adjunct model comes from L2
acquisition research, which suggests that new language is
internalized from rich content study (Brinton, Snow, 6
Wesche, 1989; Krashen, 1993; Krashen & Biber, 1988; Wesche,
1993). In the case of the adjunct model, the rich input is
drawn from the content class to which the adjunct course is
linked. A second rationale for the adjunct course draws on
the literature on learning strategy instruction (Adamson,
1991, 1993; Caverly & Orlando, 1991; Chamot, Dale, O'
Malley, & Spanos, 1992; Hernfindez, 1991; O'Malley & Chamot,
1990; Padrdn, 1992; Padrdn, Knight, & Waxman, 1986; Padrdn &
Waxman, 1988; Palincsar & Brown, 1984; Paris & Myers, 1981;
Rosenshine & Meister, 1994; Waxman, Felix, Martinez, Knight,
& Padrdn, 1995; Weinstein, 1982, 1987; Weinstein & Mayer,
1986). Strategy instruction in the Project LEAP adjunct
course includes models, peer assistance, gradual transition
from teacher- to student-control, and opportunities for
independent strategy use (Grabe, 1993; Palincsar & David,
1991; Palincsar & Klerk, 1991). Additionally, strategy
instruction is embedded in the context of the specific
subject-matter presented in the academic course.
9
Statement of the Problem
This study addresses two problems. First, all reading-
to-write tasks in the undergraduate curriculum require
summarization strategies. In writing from sources, students
need to be able to understand written text and then
distinguish what is important from what is irrelevant.
Additionally, students need to condense important
information and to transform the original text into a new
one.
It is often assumed that, by the time they enter
college, students have already acquired summarization
skills. While Ll reading research supports this assumption
(Brown, Day, & Jones, 1983; Garner, 1985; Hidi & Anderson,
1986; Winograd, 1984), L2 reading research has shown that
this may not necessarily be true for NNS (Johns, 1985; Johns
& Hayes, 1990), and that L2 students find summary writing
practice very useful (Leki & Carson, 1994; Ready & Wesche,
1992; Repath-Martos, 1992).
It has been noted that large numbers of NNS, entering
freshmen in the CSU system are "underprepared," meaning that
they test below the cut-off point for enrollment in freshman
English composition classes. These NNS, many of them from a
Spanish-speaking background, are also poor readers, as
measured by the scores they achieve on the reading skills
section of the EPT.
10
Second language reading research, similar to LI reading
research, has shown that poor readers differ from good
readers in the strategies that they use to abstract
information from written text. Poor readers use local
strategies and read in a fragmented manner, instead of
employing abstraction operations that lead to understanding
the global meaning of a text (Bensoussan & Kreindler, 1986;
Devine, 1988; Eskey & Grabe, 1988; Johns, 1991a; Johns &
Mayes, 1990). The local nature of the reading strategies
used by poor L2 and Ll summary writers can be observed in
the writers' frequent copying and in their general failure
to combine information across different sections of the
source text (Hare & Borchardt, 1984; Johns, 1985; Johns &
Mayes, 1990; Winograd, 1984).
The second area that this study addresses is related to
the quality of instruction that "underprepared" NNS receive.
As already indicated, "developmental" courses are often
isolated from those in academic departments. The
educational context in which this study is embedded attempts
to break the state of isolation in which "developmental"
instructors often find themselves. Specifically, the
integration of language and subject-matter content, which
grounds instruction in the reality of the content class
(Carson et al., 1992), finds support in L2 acquisition
research. In addition, the integration of explicit
11
cognitive instruction in summary writing into the adjunct
course provides "underprepared" native Spanish-speaking
(NSS) freshmen with academic experiences that may result in
the acquisition of academic literacy skills in English.
Examination of these two problem areas, the need for L2
summarization-strategy training and the integration of
language, content, and strategy instruction, is necessitated
for three reasons. First, the numbers of NSS students
placed in "developmental" courses are growing rapidly.
Although these students have participated in the U.S. school
system for a number of years, they are "underprepared"
because they have not yet acquired the academic English
language skills necessary to succeed in a general education
curriculum that requires high levels of academic literacy
skills in English.
Second, researchers in the bilingual education field
(Hern&ndez, 1991; Padrdn, 1992; Padr6n, Knight, & Waxman,
1986; Padrdn & Waxman, 1988) have argued that students from
a Spanish-speaking background are not participating in
school experiences that promote the acquisition of cognitive
strategies that are needed to function successfully in an
academic environment. Finally, by linking general education
courses to adjunct classes, "underprepared" L2 students can
learn how to apply newly learned English academic language
12
skills and strategies in authentic contexts (Adamson, 1991,
1993; Commander & Smith, 1995).
Purpose of the Study
The purpose of this study was two-fold. From a narrow
perspective, the purpose of this study was to determine
whether "underprepared" native Spanish-speaking, English-
speaking university freshmen would benefit from explicit,
content-based instruction in summary writing in adjunct
courses. Through the analyses of pre- and post-intervention
summary protocols, this investigation assessed the
performance of three groups of students, a group of
"underprepared" L2— NSS— freshmen concurrently enrolled in a
general education course and in an adjunct course, a group
of "underprepared" L2— NSS— freshmen enrolled in a general
education course, and a group of "able" L2— NSS-students
enrolled in the same general education course. From a
broader perspective, the objective of this study was to
determine the extent to which instructional strategies that
had been found to be successful with LI students would
provide "underprepared" L2— NSS— freshmen with specific
academic benefits.
13
Importance of the Study
Summary writing has been studied from different
perspectives. For example, Ll reading and cognitive
psychology researchers have looked at developmental
differences in the acquisition of summarization strategies
(c.f., Brown & Day, 1983: Garner, 1982, 1985; Mare &
Borchardt, 1984; Rinaudo, 1993; Solarz, 1994; Taylor, 1984)
and in task awareness (c.f., Garner, Belcher, & Smith, 1985;
Rinaudo, 1993; Winograd, 1984). In the L2 reading field, a
number of studies have focused on the summarization
strategies of foreign language (FL) students (c.f., Endres-
Niggemeyer, Waumans, 6 Yamashita, 1991; Holmes & Ramos,
1993; Kozminsky & Graetz, 1986; Sarig, 1993). By contrast,
very few attempts have been made to look at the
summarization strategies of students in L2 settings (c.f.,
Basham, 1986; Johns, 1985; Johns & Mayes, 1990), and no
studies have investigated the effect of explicit instruction
on the summarization strategies of "underprepared" NSS
students in university-level programs.
This study contributes to the literature in five ways.
First, the results of this study contribute to a small, but
growing body of literature pointing to the potential
academic benefits arising from the teaching of cognitive
strategies to Spanish-speaking students (Hernfindez, 1991;
Padrdn, 1985; Padr6n, Knight, & Waxman, 1986; Padrdn &
14
Waxman, 1988)(an argument supported by Kinsella, 1992 for
other language minority groups as well).
Second, extensive research has focused on
"international students," students who have elected to
enroll in U.S. colleges and universities (Vald6s, 1992) or
on "incipient bilinguals," immigrant students who are in the
process of acquiring English (Vald6s, 1992). In contrast,
very limited research has looked at "American bilingual
minority students," (Vald6s, 1992) students who may have
lived in this country for a long time or been born in the
U.S. (Snow, 1994). This study addresses the latter student
population, which in contrast to "international students,"
has been in the U.S. school system for longer periods of
time. However, as indicated by the Committee on Educational
Policy of the CSU system (1995), only recently have these
high school students exited programs for LEP students or
have they gained access to college preparatory courses. The
findings from this study can provide researchers and
educators with empirical evidence about the extent to which
the student population represented in this investigation can
benefit from explicit, content-based summarization-strategy
instruction.
Third, the results of this study will provide
information as to how "underprepared" NSS freshmen, through
a number of cognitive operations, condense, combine, and
IS
transform the propositions contained in a source text, prior
to, and following strategy training.
Fourth, this investigation contributes to the
identification of instructional strategies that can promote
an improvement in the reading and writing skills of
"underprepared" NNS from a Spanish-speaking background.
Fifth, the results of this study will have implications
for the language arts curricula of high school English
classes and for the curricula of university-level
developmental English courses, including programs such as
the CSU Summer Bridge, the Intensive English Experience, and
even freshman English composition classes.
Research Questions
Within the context of the purposes set forth, the
following research questions were posed,
l. To what extent, if any, do "underprepared"
university-level NSS freshmen, enrolled in one of three
general education courses— Health Science 150, Cultural
Anthropology 250, or Animal Biology 155— and receiving
explicit, content-based instruction in adjunct courses,
(Experimental *= E) differ from "underprepared"
university-level NSS freshmen, enrolled in one of the
three general education courses (Control Group 1), and
from "able," formerly "underprepared," NSS university-
16
level students, who have already completed the freshman
Composition requirement (Control Group 2), in their
understanding of a summarization task in English, as
measured by
a. the groups' holistic performance?
b. the number of main ideas reproduced?
c. the number of adjacent main ideas and the
conjunctive relationships between adjacent main
ideas?
d. the groups' summarization efficiency, as indicated
by the ratio of the number of main ideas reproduced
to the number of words included in the summaries?
2. To what extent, if any, are there differences among the
Experimental Group and Control Group 1 and control Group
2 in the strategies used to reproduce main ideas, as
measured by
a. the percentage of copies, near copies, quotations,
and paraphrases?
b. the percentage of combinations within and across
paragraphs?
Overview of the Methodology of the Study
This study involved three groups of NSS subjects
enrolled in three university-level general education
courses. The first group, called Experimental, consisted of
17
20 "underprepared" freshmen, who were concurrently enrolled
in one of three general education courses— Health Science
150, Cultural Anthropology 250, and Animal Biology 155—-and
received eight hours of supplementary summarization-strategy
instruction through their participation in an adjunct,
content-based second language instruction program. The
second group, Control Group l, was composed of 20
"underprepared" freshmen, who were enrolled in one of the
three general education courses but did not receive any
supplementary summarization instruction. The third group,
Control Group 2, consisted of ten "able" subjects— formerly
considered "underprepared"— who participated in one of the
three general education courses. As was the case with
Control Group 1 subjects, Control Group 2 subjects did not
receive supplementary instruction in summary writing.
The three groups of subjects in this investigation
participated in a pre- and post-instruction summarization
task that involved reading and summarizing one of three
content-specific articles, which contained features of
descriptive and comparative texts. The 50 subjects who
participated in this study also completed a Student Profile
Questionnaire which contained questions about the students'
general background, language and school background, and
language use.
18
Pre- and post-intervention summaries written by the
Experimental Group and by Control Groups 1 and 2 were
analyzed in three ways. First, three trained raters scored
the summaries holistically using a Holistic Rating Scale for
Summary Writing. The raters also identified the main ideas
included in the summaries. Using the Statistical Package
for the Social Sciences (SPSS) for Windows, separate
Analyses of Covariance (ANCOVAS) were conducted to test for
differences in holistic scoring and in the number of main
ideas included in the summaries written by the Experimental
Group and by Control Groups 1 and 2. In these analyses, the
dependent variables were the subjects' holistic scores on
the post-intervention task, as well as the number of main
ideas included in the post-intervention summaries. The
subjects' holistic scores on the pre-intervention
summarization task, the number of main ideas included in the
pre-intervention summaries, and the subjects' length of
residence in the United States were used as covariates.
Post-hoc ANCOVAS were run to test for statistical
differences between the Experimental Group and Control Group
1, between the Experimental Group and control Group 2, and
between Control Groups 1 and 2, in their holistic scores and
in their main idea performance.
In addition to the statistical analyses, the main ideas
contained in the 50 pre- and post-intervention summaries
19
were analyzed for combination and reproduction strategies,
and for conjunctive relationships. Combination strategies
included one idea unit reproduced individually, idea unit
combinations across sentences within paragraphs, and idea
unit combinations across paragraphs. Reproduction
strategies included instances of copies, near copies,
quotations, and paraphrases. Conjunctive relationships
included additive, adversative, causal, and temporal
conjunctions linking two adjacent main ideas.
The third type of analysis involved summarization
efficiency, as indicated by the mean ratio of the number of
main ideas included in the summary protocols to the number
of words included in the summaries. Pre- and post
intervention summarization efficiency ratios were calculated
for the Experimental Group and for Control Groups l and 2.
Limitations
The following limitations are noted:
1. The sample size was limited to a maximum of 20 subjects
in the Experimental Group, 20 subjects in Control Group
1, and 10 subjects in Control Group 2.
2. It could be argued that significant results could be due
to factors such as teacher effect or subjects'
motivation.
3. The subjects' regular attendance in the adjunct courses
20
was outside the researcher's control. Attendance was,
however, recorded.
4. Ultimate use of class time in the Cultural Anthropology
250 adjunct class was outside the researcher's control.
Definition of Terms
The key terms used in this study are defined below.
1. Content-based instruction: An approach to L2 language
instruction that integrates particular content with
language-teaching aims (Brinton, Snow, & Wesche, 1989, p.
2) .
2. Adjunct model: A model of content-based L2 instruction
in which students are enrolled concurrently in two linked
courses— a language course and a content course— with the
objective that the two courses share the content base and
complement each other in terms of mutually coordinated
assignments (Brinton, Snow, & Wesche, 1989, p. 16).
3. Explicit instruction: A method of instruction which
involves explicitly outlining component processes of a
desired strategy and developing lessons to teach the
components (Hare & Borchardt, 1983). Explicit instruction
in this study includes explicit models and guided and
independent practice (Guido & Colwell, 1987) in a
cooperative context (Grabe, 1993).
21
4. Able students: Students who participated in a minimum
of one pre-baccalaureate reading/writing class and, at the
time of this investigation, had'completed the CSU Freshman
English composition requirement.
5. Underprepared students: Students who, upon admission to
the CSU system, were required to enroll in a minimum of one
pre-baccalaureate reading/writing class, and at the time of
this investigation, had not fulfilled the requirement.
6. Limited English proficient (LEP): A subgroup of
language minority/non-English language background [students]
whose English proficiency was not sufficient for them to
participate effectively in an English-only classroom
(Macias, 1993, p. 232).
7. Non-native speakers of English (NNS): Students whose
first language is not English.
8. Native Spanish-Speaking (NSS): Students whose first
language is Spanish.
9. Latino; All persons generally within the United States
whose origins or ancestry can be traced to one of the
Spanish-dominant speaking Latin American countries, the
Caribbean nations of Puerto Rico, Cuba, and the Dominican
Republic (Macias, 1993, p. 255).
10. Freshman: A term indicating that a student has been
admitted to study in a college or university and has
22
enrolled in and is in the process of taking and/or has
completed 1 to 45 quarter units of academic credit.
11. Sophomore: A term indicating that a student has been
admitted to study in a college or university and has
enrolled in and is in the process of taking and/or has
completed 46 to 90 quarter units of academic credit.
12. Junior: A term indicating that a student has been
admitted to study in a college or university and has
enrolled in and is in the process of taking and/or has
completed 91 to 135 quarter units of academic credit.
Organization of the Remainder of this Dissertation
Chapter II provides a review of the literature relevant
to this investigation. Specifically, the chapter presents
research on the following topics: the constructive nature
of reading-to-write tasks, summary writing in the LI and L2
reading fields, text structure and linguistic cohesion in
text, explicit instruction, content-based second language
instruction including the adjunct model, and previous
research related to the student population represented in
this study. Chapter III provides a description of the
subjects who participated in this study and the nature of
the intervention, chapter III also presents a description
of the data collection procedures, as well as a discussion
of the variables and the methods employed to analyze them.
23
The findings of the study are presented in Chapter IV
and a discussion of the findings is included in Chapter V.
Chapter VI provides a summary of the discussion, the
conclusions of the study, and recommendations for further
research and classroom pedagogy.
24
CHAPTER II
REVIEW OP PERTINENT LITERATURE
This chapter presents research and theories which are
directly relevant to this study and support the design and
methodological decisions found in preceding and succeeding
chapters. This chapter presents a review of the literature
in seven distinct areas: (1) the constructive nature of
reading-to-write tasks; (2) summary writing, including the
cognitive operations involved, and first language (LI) and
second language (L2) research in summary writing; (3) text
structure and empirical research on text structure; (4)
linguistic cohesion in text; (5) the role of explicit
instruction in cognitive psychology and applied linguistics;
(6) content-based L2 instruction and relevant research on
the adjunct model; and (7) previous studies related to the
specific student population represented in this study. The
chapter concludes with a summary of the seven areas covered
in this review.
Readina-to-Write as a Constructive Act
Current research on the psycholinguistics of reading
argues that readers construct meaning by connecting and
integrating old knowledge with new information (Ackerman,
1989). In this view, the meaning of a text does not lie in
the text itself (Bartlett, 1932; Rumelhart & Ortony, 1977);
25
rather, text comprehension is the result of an interactive
process between the text and the reader's background
knowledge (Carrell & Eisterhold, 1988; Eskey, 1988; Kucer,
1985; Rumelhart, 1977a, 1980).
Much like reading, composing is a constructive act.
For example, Flower (1987, 1989, 1990) and Flower and Hayes
(1980) hypothesize that meaning resides in the writer's mind
in the form of multiple representations. For Flower (1987,
1990), task representation is an interpretive process which
is influenced by the writer's prior knowledge and
experience, the format and features of the text, the
writer's strategies for composing, and the writer's goals
(see Figure 1).
I M K INVIRIINMINr
l l l l ft HI IllH II AL
fftllllUM
TifH
A in lm r
liw m f
T lftT
pftiiori.rD
to FAR
WHITING H U U I S U
T H I WRITER'S LONG-TTRM
MEMORY
TRANSLATING
IfwiMwlftr Tiip».
RtVISINO
Figure l. Flower and Hayes' structure of the writing model.
26
If both reading and writing can be understood as
constructive processes, then reading-to-write can also be
conceived as a constructive act (Ackerman, 1989; Spivey,
1987). Reading-to-write tasks may involve the use of only
one source (e.g., in summarization tasks) or more than one
source (e.g., in discourse synthesis tasks). Critical to
the constructive nature of reading-to-write tasks is the
fact that readers create new texts (Ackerman, 1989; Nielsen,
1995; Spivey, 1983, 1987, 1990; Spivey & King, 1989).
The degree of difficulty of a reading-to-write task
does not solely depend on the nature of the task, but is the
result of a number of interrelated factors. Background
knowledge, the degree of complexity of the source text the
reader is attempting to transform, and the level of
invention required to complete the task affect the reader
(Flower, 1990). For example, summarizing a number of
related but internally complex papers in different
disciplines may be a more difficult task than synthesizing
sources by selecting information closely tied to the
students' background knowledge (Flower, 1990).
For Spivey and King (1989), the construction of meaning
is affected by factors within the reader, such as maturity
and reading ability, and by the nature of the reading task.
For example, Spivey (1983) found that reading ability
affected the way in which her subjects selected, connected,
27
and organized information. In her research, Spivey found
that proficient readers at the college level were more
sensitive to the importance of textual ideas and produced
texts that included a higher percentage of important
information than did those produced by less able readers.
When compared to the texts produced by less able readers,
the able readers' texts were more integrated, compact, and
better organized.
The notion that the nature of the task affects the
reader-writer has been confirmed in a number of studies.
For example, process research which compared the performance
of students on reading-to-write tasks revealed that analytic
writing, as opposed to summary writing, led students to
think more deeply (Langer & Applebee, 1987), to attend to
their writing process, and to make more abstract
interpretations (Durst, 1984). In another study on reading-
to-write tasks, Kantz (1987) found that students who
approached a writing task as summarizers did little
manipulation of the source text as opposed to students who
saw themselves as explicators. In her analysis of the
behaviors or more and less fluent college readers engaged in
reading-to-write tasks, Kennedy (1985) found that, in
contrast to less proficient readers, fluent readers planned
actively and interacted with the source texts more often.
28
From this discussion, it can be inferred that knowledge
telling and knowledge transforming are not functions of a
source text. Knowledge telling and transformation occur in
the mind of the reader; they refer to mental processes by
which new texts are created (Bereiter & Scardamalia, 1987;
Flower, 1990).
As shown in the first section of this chapter, summary
writing is a distinctive reading-to-write task. The next
section presents a review of the literature related to the
cognitive operations involved in summary writing, as well as
the cognitive demands of reader- and writer-based summaries.
Finally, the section presents empirical research in summary
writing in the fields of cognitive psychology and Ll and L2
reading and writing.
Summary Writing
Cognitive Operations
Rumelhart (1977b) argues that a text is defined by its
constituent structure: the major constituents of the text,
central to a story, are represented directly in the highest
level of the text. Conversely, the story details, low in
importance, are peripheral in the structure of the story.
Rumelhart's model of summarization focuses on the text
and its constituent structure. For Rumelhart, text
29
summarization is a process that requires the reader to trim
the tree structure so that nodes below the chosen level in
the tree are removed. The decision about which nodes to
trim is determined by how much detail is required in the
summary and how much information is presented in the node in
question (Carson, 1993).
According to Kintsch and van Dijk (1978; van Dijk
(1980; van Dijk & Kintsch, 1983), the semantic structure of
a text can be accounted for in terms of macrostructures and
microstructures. Macrostructures (van Dijk, 1980) "account
for the various notions of global meaning, such as topic,
theme, or gist" (p. 10). The microstructure of the text is
the "local level of the discourse, that is, the structure of
the individual propositions and their relations" (Kintsch &
van Dijk, 1978, p. 365). Specifically, in his discussion of
microstructures, van Dijk (1980) refers to words, phrases,
clauses, sentences, and connections between sentences.
Kintsch and van Dijk (1978) argue that, for a text to
be coherent, propositions have to be connected, not only at
the local level, but also at the level of semantic
macrostructures. This means that a text is coherent only if
the propositions in the text are linked both at the local
and the global levels, and macrostructures are needed for
those relations to be made explicit.
30
Kintsch and van Dijk's (1978) summarization model
focuses on mental operations. In their model, the gist of a
text can be generated by applying the following macrorules:
1. Deletion. Each proposition that is
neither a direct nor an indirect
interpretation condition of a subsequent
proposition may be deleted.
2. Generalization. Each sequence of
propositions may be substituted by the
general proposition denoting an immediate
subset.
3. Construction. Each sequence of
propositions may be substituted by a
proposition denoting a global fact of which
the propositions denoted by the
microstructure propositions are normal
conditions, components, or consequences.
(p. 366)
As van Dijk (1980) explains, when readers apply the
first macrorule, deletion, they engage in a process of
deleting propositions which are not needed for the
interpretation of other propositions in the text. The
deletion rule can be understood as a selection rule,
requiring the reader to select from the text those
propositions that act as presuppositions of other
propositions. When the generalization rule (Macrorule 2) is
applied, individual propositions are subsumed into more
general propositions. The third macrorule, construction,
requires readers to construct a new proposition at a more
global level.
31
The rules are reductive, constructive, and
organizational (van Dijk, 1980). They are reductive because
semantic information is lost. They are constructive because
certain elements can be combined into new, more complex
propositions (Spivey, 1983)• Finally, the macrorules are
organizational because whole sequences of propositions can
be linked. As explained by Spivey (1982), "the macrorules
can also be applied recursively until all that remains is
the most thematic proposition of the text" (p. 46).
Spivey (1990) proposes a model of summarization which
also includes three operations: organizing, selecting, and
connecting. In Spivey's model, the three operations
involved in summarizing do not differ from those required to
write an argument, a proposal, or any other kind of text.
For Spivey, organizing involves developing a text structure
different from the source text, although organizing, for
Spivey, does not occur when summary writers produce
isophormic summaries which replicate the structural pattern
of a source text. Selecting involves identifying important
information that should be included in the summary; in this
selection process, summary writers use textual relevance as
a means to help them decide which information is redundant
or irrelevant. Connecting involves making connections
between background knowledge and newly acquired knowledge.
Spivey's contention is that although summaries are not meant
32
to include the writers' elaborations on the source text,
good summary writers will make inferences (Brown & Day,
1983; Kintsch & van Dijk, 1978).
Drawing on the analyses of Kintsch and van Dijk's
(1978) macrorules for text summarization and of protocols
obtained from children and adults, Brown and Day (1983)
identify six basic rules for summarization. The first two
rules, equivalent to Kintsch and van Dijk's deletion
macrorule, involve deleting trivial and redundant
information. The third and fourth rules, contained in
Kintsch and van Dijk's generalization macrorule, involve
substituting a superordinate action for a list of items or
actions, and substituting a superordinate action tor
subcomponents of that action. The remaining two rules,
construction in Kintsch and van Dijk's model, involve
selecting a topic sentence tor the paragraph or inventing a
topic sentence, if the paragraph does not include one.
Classification of Summaries
Writing is a social activity which involves inferring a
reader's thoughts and feelings in an act of communication.
In this context, writing for an audience implies
decentering, or avoiding egocentric tendencies (Piaget cited
in Flower, 1979, p. 20) or inner speech (Vygotsky cited in
Flower, 1979, p. 20). Both inner speech and egocentrism
33
mean being centered in one'e own thoughts and not being able
to take into account the readers' perspective (Kaplan, 1988;
Maimon, 1979; Perl', 1979; Rafoth, 1985; Ushakova, 1994).
While the most salient characteristics of inner speech
or egocentrism tend to disappear with age, it is the case
that novice college writers still frequently fail to take
into account the reader's perspective. This failure to
transform inner thought into reader-based prose is what
results in writer-based prose (Flower, 1979). It assumes a
high degree of shared knowledge, leaving it up to the reader
to understand what is being said (Hinds, 1983, 1987). This
practice may be attributed to the students' inability to
move from writer- to reader-based prose (Flower, 1979) and
their lack of familiarity with certain academic writing
tasks (Ackerman, 1990; Flower, 1990; Horowitz, 1986; Rose,
1983) .
Drawing on Flower's (1979) notion of reader- and
writer-based prose, Hidi and Anderson (1986) make a
distinction between reader- and writer-based summaries.
Hidi and Anderson's categorization of summaries, however,
takes into account the purposes for which summaries are
written, instead of whether or not the writers' are aware of
their audience. According to Hidi and Anderson, summaries
written for the purposes of being read by the writer only
are writer-based; by contrast, summaries written for the
34
purposes of being read by an audience other than the writer
are reader-based.
Because writer-based summaries are meant to be read by
the writer only, syntax may be reduced to a minimum and the
writer does not need to worry about grammar rules, sentence
cohesion, or space limitations (Carson, 1993; Hidi &
Anderson, 1986). A study summary, usually meant to provide
the writer with a record of important information, is an
example of a writer-based summary.
Reader-based summaries are usually written for the
benefit of an audience, typically a content instructor (Hidi
& Anderson, 1986), and are usually produced for the purposes
of completing a course requirement. As Hidi and Anderson
(1986) argue,
...the function of this type of summary
varies from having to demonstrate one's
ability to pull out and condense accurately
the important ideas of a given text, to
trying to arouse people's interest by
providing them with the gist of some written
text, as in the case of a newspaper's book
critic who is recommending a book to
readers. (p. 479)
The characteristics of reader-based summaries suggest
that grammatical accuracy, sentence cohesion, space
limitations, and the nature of the summary to be produced
are important constraints for the summary writer (Hidi &
Anderson, 1986) and may increase the writer's cognitive load
(Kirkland & Saunders, 1991).
35
Empirical Research in Summarv_Wrltlnq
As already noted in Chapter I, the ability to summarize
involves the comprehension, condensation, and transformation
of important ideas at the expense of less important
information. This section presents current research in
summary writing in the fields of cognitive psychology and LI
and L2 reading1.
DevelopmentaL_differences_inthe summarization-strateav
acquisition process. Extensive research on age differences
and summarization strategies in the fields of cognitive
psychology and LI reading has led investigators to conclude
that summarization strategies develop slowly over time
(Brown & Day, 1983; Garner, 1982, 1985; Garner & McCaleb,
1985; Hare & Borchardt, 1984; Johnson, 1983; Ricciardi,
1984; Rinehart 6 Thomas, 1993; Taylor, 1986).
According to Garner (in Hidi & Anderson, 1986), there
are three stages in the summarization-strategy acquisition
process. The first stage, deficiency, involves the non
application of any summarization strategy. In this stage,
summary writers do not distinguish important from
unimportant information and, instead, confuse interesting
information with information that has textual relevance.
The second stage, inefficiency, is characterized by the use
of the copy-delete (Brown & Day, 1983) strategy, the first
strategy to be acquired (Garner in Hidi & Anderson, 1986).
36
In the last stage, efficiency, summary writers consistently
employ sorting strategies, 'identify important information,
and restructure text.
Developmental differences in task awareness and
sensitivity to importance. There is evidence that supports
the idea that children understand the requirements of a
summarization task. For example, the summaries produced by
children in first- (Johnson, 1983) and fourth-grades (Hidi,
1985) were found to be shorter than their recalls. Hidi and
Anderson (1986) concluded that these findings suggest that
young children are aware that summarization involves the
reduction of a source text.
Research has also shown that young children, including
students in fifth-grade (Garner, Belcher, Winfield, & Smith,
1985), eighth-grade (Winograd, 1984), and ninth- and
eleventh-grades (Garner, 1985) can verbalize the
requirements of a summarization task.
The ability to identify central information and to
produce it in a summary characterizes good readers and
develops over time and with experience (Garner, 1985). For
example, Hare and Borchardt (1984) found that high school
students did not identify implicit topic sentences. In his
study of eighth-grade and adult readers, Winograd (1984)
found that some poor readers failed to identify information
that the adults and good readers considered important.
37
Garner (1985) reached similar conclusions in her analysis of
the performance of children and adult readers. In her
study, Garner found that, when compared to ninth- and
eleventh-grade students, college students reproduced nearly
all of the important ideas in the source text. Garner's
findings, however, were contradicted by Rinaudo (1993), who
noted a deficiency in the ability of college students to
identify important ideas in a source text. Fifty per cent
of Rinaudo's subjects were able to identify only 50% of the
important information in a source text. Likewise, Strode
(1989) reported that college students receiving annotation
strategy training did not produce a higher number of key
ideas on a summarization task.
The developmental nature of condensation, and the
ability to create tonic sentences. The ability to combine
information, make generalizations, and include superordinate
concepts appears to be a characteristic of mature summary
writers (Hidi & Anderson, 1986). For example, Brown and Day
(1983) found that, when compared to children, adolescents
and college-level summarizers were able to combine and
rearrange information across paragraphs.
Brown and Day (1983), Brown, Day, and Jones (1983), and
Winograd (1984) observed that, when compared with young
children, older summary writers condensed more ideas in the
same number of words. These observations were not supported
38
by Garner (1985), whose college subjects produced less
succinct summaries than did ninth- and eleventh-grade
students.
Research by Brown and Day (1983) revealed that graduate
students, considered expert summary writers, created topic
sentences and built their summaries around them. By
contrast, non-expert college students (Brown & Day, 1983;
Garner & McCaleb, 1985) invented topic sentences in only
about half of the instances that such sentences were
expected.
According to Taylor (1984), professional writers differ
from inexperienced high school and college students in their
summarization processes. In his study, Taylor found that
professional writers studied the text more carefully,
planned more, remained more objective, and took the audience
into account. In another study of expert writers, Kaplan,
Cantor, Hagstrom, Kamhi-Stein, Shiotani and Zimmerman (1994)
examined the abstracts submitted to a professional
organization. Kaplan et al. found that the abstracts were
topic-based and detached, rather than interactional and
involved, and they contained fully elaborated syntactic
structures.
Summarization-strateav training. Research has revealed
that summarization instruction is effective in the
elimination of details and redundant ideas (Charry, 1987;
39
Day, 1980; Hare & Borchardt, 1984; Kin?, 1992; Ko2nlnsky &
Graetz, 1986; Maxworthy & Barry, 1992; Strode 1989; Taylor &
Beach, 1984). Summarization instruction, however, does not
seem to result in the construction of macrostructures (Day,
1980; Hare & Borchardt, 1984).
Second Language Research and Summary Writing
A growing body of research has focused on L2 students
and their summarization strategies. This section presents a
summary of the literature related to students in L2
contexts.
Basham (1986) analyzed the summaries produced by
inexperienced Alaska Native writers enrolled in college-
level courses. In her study, Basham found that the
students' summaries were characterized by the use of oral,
informal language instead of written, formal language.
In an earlier but related study, Johns (1985)
investigated the summarization strategies of underprepared
minority students enrolled in university-level remedial
reading and writing classes. Johns found that, when
compared to adept students, underprepared students omitted a
number of main ideas from the source text, produced fewer
combinations of idea units, and used more words to produce
less information. These results were later confirmed in an
investigation by Johns and Mayes (1990). In this study,
40
Johns and Mayes (1990) compared the summarization strategies
of a group of low-proficiency students enrolled in remedial
English as a second language (ESL) reading courses and those
of a group of high-proficiency ESL students who, at the time
of admission to the university, had passed the English
Placement Test and, at the time of the study, had fulfilled
the freshman English composition requirement. Johns and
Mayes found that the low-proficiency ESL students in
college-level courses produced more replications or direct
copies from the source text. Low- and high-proficiency ESL
students were equally unable to condense ideas within
paragraphs, to combine ideas across paragraphs, and to
produce macropropositions.
Second language proficiency was also found to be an
important factor in a study by Campbell (1990). Campbell
analyzed how native English-speaking and less- and more-
proficient non-native university students used information
from a background text. In her study, Campbell noted a
developmental trend in the proportions of paraphrases and
summaries used by students in the three groups. The three
groups differed in the holistic scores received, with the
native group receiving the highest mean scores and the least
proficient non-native group receiving the lowest mean
■scores. The developmental nature of the ability to use
written sources had been noted in an earlier study (Britton,
41
Burguess, Martin, McLeod, & Rose, 1978). According to
Britton et al., a writer nay have to go through stages that
include suranarizing, expanding written ideas, and imitating
style, until he or she reaches the synthesizing stage.
Campbell's results regarding the effect of language
proficiency when writing from sources were confirmed by
Connor and Kramer (1995). In their case studies, Connor and
Kramer found that the performance of students with low ESL
proficiency, as measured by the TOEFL test, was affected by
the students' limited vocabulary, and poor grammar, syntax,
and discourse**level reading and writing strategies.
The fragmented nature of the reading process of L2
readers was pointed out in a study by Kozminsky and Graetz
(1986), who found that although advanced English as a
foreign language (EFL) students employed a Jbottora-iip
approach to studying written text, their summaries were
constructed top-down, meaning that the readers relied on
their general knowledge.
Another line of research, represented here by three
studies, argues that summarization strategies transfer
across languages. First, Cumming (1989a) investigated the
thinking process of 14 French- and English-speaking students
who performed reading and summarizing tasks in Ll and L2.
Cumming found that the students used equivalent strategies
in both languages and the use of problem-solving strategies
42
did not seen to be related to the subjects' L2 proficiency.
The transfer of strategies from LI to L2 was confirmed"by
Sarig (1993), in a qualitative analysis of two study
summaries— one in Hebrew and the other in English-~prepared
by an advanced EFL student. Sarig found that the summary
products in both languages were closely related and that,
although the student did not produce macrostructures, his
summarization process included metacognitive activities and
clarify, link, transform, and revise activities.
Similar summarization processes in Ll and L2 were
identified in a study by Endres-Niggemeyer, Waumans, &
Yamashita (1991). These researchers reported that one
speaker of German as a native language and two speakers of
German as a foreign language (FL) did not differ in their
summarization processes. Specifically, the researchers
concluded that the readers engaged in a process of
"inferencing, planning and control, knowledge acquisition,
relevance assessment, meaning reduction, condensation,
construction, and output" (p. 523).
The findings of the three qualitative studies were
further confirmed in a study of ESL writers (Cumming,
1989b). Cumming examined the performance of 23 L2 writers
on three writing tasks— a letter, a summary, and an
argument— in relation to the subjects' writing expertise and
L2 proficiency. Cumming found that experienced Ll writers
43
produced more effective discourse organization and content,
and interrelated planning and production processes in
sophisticated ways. Cumming also found that L2 proficiency
affected overall writing quality.
The testing situation is a factor that has been found
to affect L2 students in their performance on a
summarization task. For example, Cohen (1993) found that
guided instructions seemed to have a mixed effect on the
summaries produced by native speakers of Hebrew and a
positive effect on the EFL students' summaries. Holmes and
Ramos (1993) analyzed the study summaries produced by FL
students enrolled in an English for Specific Purposes (ESP)
course under an experimental condition, consisting of a time
limit, and a control condition. Holmes and Ramos reported
that, under the experimental condition, the summaries were
not carefully organized, included fewer metastatements and
more copies, and omitted information which did not appear in
the first sections of the source text. By contrast, the
summaries produced under the control condition reflected the
structure of the original text, included more metastatements
and paraphrases, and omitted less important information.
Although the studies focusing on the effect of
instruction on L2 students are still limited in number,
their results confirm prior Ll findings. For example, in a
study in the FL context, Cordero-Ponce (1994) trained
44
intermediate-level French as a FL students in the use of
Mare and Borchardt*s summarization rules. Post-test results
showed that the experimental group outperformed the control
group in the number of idea units recalled on a summary-
writing task in English. In another study in the EFL
context, Bensoussan and Kreindler (1990) instructed their
students in the use of Kintsch and van Dijk's macrorules.
Bensoussan and Kreindler (1990) concluded that instruction
in the macrorules helped students grasp the gist of the
source text. Initially, however, students "were frustrated
by their inability to distinguish macrostructures from
microstructures'1 (p. 65).
As already noted, research has shown that L2
proficiency affects the readers' ability to summarize
information from a given text and that Ll and L2 students
can benefit from summarization instruction. However, the
structure of the source text can constrain the readers'
ability to comprehend and recall information. Therefore,
the following section presents a review of the literature
related to text structure and Ll and L2 students.
45
Text Structure
Drawing upon Grimes (1975), Heyer (1975, 1981, 1982)
has gathered evidence for five basic ways to organize
discourse: collection, causation, problem/solution,
comparison, and description:
1. collection: relation which shows how
ideas or events are related into a group on
the basis of some commonality;
2. causal: relation which shows a causal
relationship between ideas where one idea is
the antecedent or cause and the other is the
consequence of effect;
3. response: a problem and solution (also
remark and reply, question and answer
formats) is similar to the causal in that
the problem is an antecedent for the
solution. However, in addition there must
be some overlap in topic content between the
problem and the solution; that is, at least
part of the solution must match an aspect of
the problem;
4. comparison: points out differences and
similarities between two or more topics;
5. descriptive: gives more information
about a topic by presenting attributes,
specifics, manners, or settings. (Heyer,
1981, p. 8)
As explained by Heyer (1982) and Heyer and Freedle
(1984), most texts will include more than one of the five
basic types. In addition, the collection and description
structures often combine to form a collection of
descriptions, which includes a collection of attributes,
settings, and specifics about a topic (Heyer 6 Freedle,
1984) .
46
Empirical Research_on the Effect of Text Structure
Several studies have shown that readers who can
identify and use the rhetorical structure of the original
passage recall more information (Bartlett, 1978; Meyer,
1979, 1982; Meyer, Brandt, & Bluth, 1980). Text structure
research has shown that different text structures can be
more or less effective in facilitating information recall.
For example, texts containing the collection of description
and description structures have been found to be the least
effective in facilitating information recall (Carrell, 1992;
Meyer, Brandth, & Bluth, 1980; Meyer & Freedle, 1984).
Moreover, text structure research has shown that
information located at the top level in the content
structure of a passage can be recalled better than
information which is at lower levels (Kintsch & van Dijk,
1978; Meyer, 1977; Meyer & McConkie, 1973).
Signaling is another factor that affects text recall
(Geva, 1992; Pearson and Camperell, 1994). For example,
Meyer, Brandt, and Bluth (1981) found that the presence of
connectives facilitated recall among ninth-grade poor
comprehenders but did not among skilled readers. A study by
Geva and Ryan (1985) revealed that when conjunctions were
missing from the text, more skilled fifth- and seventh-grade
readers inferred the missing links; in contrast, less
skilled readers did not. In their study, Geva and Ryan
47
found that readers at all levels improved their
comprehension of intersententlal relations when their
attention was directed explicitly to the conjunctions in the
text.
Training in the recognition of rhetorical structure has
resulted both in improved recall of text and in improved
quality of expository text (Armbuster & Anderson, 1980;
Armbuster, Anderson, & Ostertag, 1987; Barnett, 1984;
Bartlett, 1978; Berkowitz, 1986; Boyle & Peregoy, 1990:
Brookes & Danserau, 1983; Geva, 1983; Gordon, 1990; Guri-
Rozenblit, 1989; Slater, Graves, & Piche, 1985; Slater,
Graves, Scott, & Redd-Boyd, 1988; Taylor, 1982; Taylor &
Beach, 1984).
Much of the research about L2 students and text
structure has been conducted by Carrell (1984a, 1984b, 1985;
Carrell, Pharis, & Liberto, 1989). Carrell's work, and more
recently Raymond's (1993) findings, have shown that explicit
instruction in text structure facilitates ESL reading.
Additional research on text structure has been
conducted by Connor (1984) and Connor and Mccagg (1987). In
her 1984 study, Connor found that native English speakers
outperformed ESL students and native speakers of Japanese
and Spanish in total recall, but she did not identify
differences for the recall of high-level ideas. The same
conclusions were reached by Connor and McCagg (1987) in
48
their investigation of a recall task. Connor and McCagg
found that ESL students and native English speakers did not
differ in the recall of important information, but the ESL
group recalled fewer subordinate ideas and was constrained
by the organization of the original passage. In an analysis
of the students' paraphrases, Connor and McCagg (1987)
concluded that the highest rated paraphrases sounded
"scientific and objective" (p. 80), began with a proper
explication of the pragmatic condition, and included
sophisticated discourse markers.
One of the variables investigated in this study was the
type of conjunctive relations between two adjacent main
ideas present in the subjects' summaries. The following
section provides an introduction to linguistic cohesion in
text, as well as empirical research regarding L2 writers and
text cohesion.
Linguistic Cohesion in Text
"Linguistic cohesion offers the means whereby elements
that are structurally unrelated to one another are linked
together, through the dependence of one on the other for its
interpretation" (Halliday & Hasan, 1976, p. 27). In
analyzing the cohesion of a given text, Halliday and Hasan
focus on across sentence ties. Ties are characterized by
49
their direction, their distance, and their type. The
direction of a tie is anaphoric to a preceding element,
cataphoric to an element which will be stated next in the
text, or exophoric to an element outside the text. Distance
can be immediate, when it occurs between items in
consecutive sentences, mediated when the reader has "to go
through another element before reaching the related item
(Spivey, 1983, p. 31), or remote, when the link does not
occur in consecutive sentences. As for type, Halliday and
Hasan*s taxonomy includes reference, substitution, ellipsis,
conjunction, and lexical.
Several researchers analyzed the use of cohesion
devices in student writing. For example, Witte and Faigley
(1981) found that more proficient writers used more cohesive
ties than less experienced writers. They also found that
the compositions produced by more proficient writers tended
to contain more immediate and mediated ties than the
compositions produced by less proficient writers, which in
turn contained more remote and mediated-remote ties. In
another related study that focused on the compositions of
Spanish, Arabic, Chinese, and Japanese students of ESL,
Ferris (1992) found that, when writing in English, more
proficient ESL students used more cohesive ties than did
less proficient students. Finally, Montafto-Harmon (1988)
reported that the compositions written in Spanish by
SO
Mexican-Spanish writers were characterized by the use of
additive conjunctions and the compositions written in
English by Anglo-American students were characterized by the
pattern of enumeration.
Because the subjects in this study received
summarization training through explicit instruction, the
following section presents a review of research and thought
in the area of explicit instruction in the fields of
cognitive psychology and applied linguistics.
Explicit Instruction
General cognitive strategy instruction is an approach
to teaching strategies that includes explicit instruction in
cognitive strategies, teacher guidance, and gradual
withdrawal of teacher support.'' In general cognitive
strategy instruction, models, guided practice, checklists,
and think-alouds are used to help students (Brown, Campione,
& Day, 1981; Brown & Palincsar, 1989; Casazza, 1993; Grabe,
1993; Guido & Colwell, 1987; Hare & Borchardt, 1984;
Palincsar & David, 1991; Rosenshine & Heister, 1994).
Statistically significant differences have been
identified for students who were taught to use cognitive
strategies through explicit instruction (Bereiter & Bird,
1985; Brown, Palincsar, & Armbuster, 1994; Davey & McBride,
SI
1986; Raphael & Pearson, 1985). Successful models of
explicit Instruction include the teaching of how, when, and
why the strategies should be used (Armbuster & Brown, 1984;
Clark, 1990; Derry & Murphy, 1986; Weinstein & Mayer, 1986).
Explicit Instruction, also known as informed training,
explicit provision of metamemory (Pressley, Levin, &
Ghatala, 1984), and direct instruction (Hare & Borchardt,
1984), can be contrasted with blind training (Campione,
Brown, & Ferrara, 1982). In blind training, students are
induced to use tactics without being told that the
strategies are appropriate for certain tasks or that the
strategies will help the students' performance (Derry &
Murphy, 1986).
Much as in the field of cognitive psychology, the role
of consciousness in learning is central in the field of
applied linguistics. A recent issue of the international
Association of Applied Linguistics Review (Hulstijn &
Schmidt, 1994) was devoted to consciousness in L2 learning.
The issue editors, J. Hulstijn and R. Schmidt, indicate that
there is a growing appreciation of the role of consciousness
in L2 learning. For example, Hulstijn and Schmidt argue
that "explicit instruction is more likely to facilitate L2
acquisition in the case of some features of language than in
others" (p. 9). The volume contributors agree that, in L2
learning, explicit instruction may play the role of a
52
"cognitive focusing device for learner attention" (p.9), a
position shared by VanPatten (1994) and Harley (1994) in the
volume and— though with variations— by a number of applied
linguists (Doughty, 1991; Ellis, 1990, 1993; Schmidt, 1990,
1992; Sharwood Smith, 1981; Terrell, 1991).
The subjects in this study learned how to summarize in
an adjunct, language-through-content course. For this
reason, the following section presents an overview of
content-based L2 instruction, and provides empirical
research on the adjunct model.
Content-Based Second Language Instruction
Traditional approaches to L2 instruction are primarily
structure- or language-based; instead, in content-based
approaches, language is the medium of instruction (Mohan,
1986; Snow, 1993). In content-based instruction, the
content around which a course is built provides students
with a critical mass (Grabe, 1986) of information that will
make the subject comprehensible (Eskey, 1992).
Content-based instruction emphasizes incidental
language learning; that is, new language is internalized
from rich content input (Krashen & Biber, 1988; Wesche,
1993). Support for content-based instruction comes from the
literature on the Canadian immersion programs (Gee, 1987;
53
Lambert & Tucker, 1972), bilingual education (Collier, 1987,
1989; Cummins, 1981, 1989), reading instruction (i.e., In-
depth reading, Dubin, 1986; narrow reading, Krashen, 1981),
and ESP (Johns, 1991b, 1992; Johns & Dudley-Evans, 1991).
In their review of content-based instruction models,
Brinton, Snow, and Wesche (1989) and Wesche (1993) identify
three models: theme-Jbased, sheltered, and adjunct. In the
theme-based model, "the language class is structured around
topics or themes, with the topics forming the backbone of
the course curriculum" (p. 14).
In the sheltered model, "a content curriculum is
adapted to accommodate the students' limited proficiency in
the language of instruction" (Crandall, 1992, p. 115).
Sheltered instruction is widely used in immersion and two
way/bilingual immersion programs, and has been adapted for
use in programs with large numbers of intermediate- or
advanced-level limited English proficient (LEP) students.
In the third model, adjunct,
...students are enrolled concurrently in two
linked courses— a language course and a
content course— with the idea being that the
two courses share the content base and
complement each other in terms of mutually
coordinated assignments. Second language
learners are sheltered in the language
course and integrated in the content course,
where both English and nonnative English-
speaking students attend the same lecture.
(Brinton, Snow, & Wesche, 1989, p. 16)
54
The adjunct model is highly dependent on the
coordination between the content and the language instructor
(Brinton, Snow, & Wesche, 1989; Crandall, 1992; Gee, 1992;
Goldstein, Campbell, & Cummings, 1994). ESL teachers are
responsible for providing instruction in the language skills
that are necessary for the students to perform successfully
in the content class (Brinton, Snow, & Wesche, 1989; Master,
1992); and content instructors are responsible for teaching
the discipline (Benesch, 1988) and for enhancing their
teaching strategies to meet the academic language needs of
L2 students (Snow, 1994). Linking content courses with
language courses provides content and language instructors
with opportunities to strengthen the curriculum and to find
solutions to curricular concerns (Benesch, 1987).
The adjunct model can take several forms. For example,
at California State University, Los Angeles (CSLA),
undergraduate students are concurrently enrolled in a
content class and in a credit-bearing adjunct class. The
adjunct class is team-taught by a language specialist and a
study group leader with knowledge about the class content.
Adamson (1990, 1991, 1993) proposes a modified adjunct
model, whereby students attend a content and an adjunct
class but do not receive credit for the ESL course.
Adamson's model integrates the use of authentic text and
explicit instruction in academic strategies in "a language
55
through content course" (1991, p. 69). In this model,
teaching is interactive, experiential, and focuses on
content areas that students will need to know when they
enroll in academic courses.
The Adjunct Model; Empirical Research
Crandall (1992) observes that research aimed at the
evaluation of the integration of language and content is
still limited, is very much needed, and, according to Wesche
(1993), is just being undertaken. This section presents a
number of studies that focused exclusively on the adjunct
model.
Several researchers have shown the effectiveness of the
adjunct model in helping students improve both their
perceptions of their academic skills and their actual
performance. For example, in a retrospective program
evaluation of the adjunct program at the University of
California, Los Angeles (UCLA), Snow and Brinton (1988)
reported that former program students felt that they could
read and write better as a result of the program.
Specifically, the academic skills that students found most
helpful included note-tedcing and prewriting, proofreading
tor errors in written work, and preparing reading
guides/notes. In addition, the results of a simulated final
examination revealed that, although the adjunct program
56
students had attained significantly lower placement scores
than those of a control group, no significant differences in
the performance between the two groups were identified at
the end.
Zn another investigation, Tricamo and Snow (1994)
evaluated the performance of students participating in
Project LEAP: Learning-for-Academic-Purposes— the adjunct
program at CSLA— with that of non-project LEAP students. In
Project LEAP, nine general education courses were linked to
adjunct courses. In their classes, general education
instructors enhanced their instructional strategies. For
example, content instructors revised their syllabi to make
expectations clear, they redesigned their paper assignments
to attend to the writing and research processes, and they
improved their lecturing strategies (Snow, 1994). The
adjunct classes— team-taught by language specialists and
study group leaders— provided students with support in
specific content areas and helped students enhance or
develop the academic strategies needed to fulfill the
requirements of the content courses. Tricamo and Snow
(1994) found that, although the adjunct program included a
higher number of special admission students than those
enrolled in the content courses, Project LEAP participants
in six of the nine courses received a smaller percentage of
57
unsatisfactory course grades (D, F, W, U, I) than non-
Project LEAP students.
In a study of the adjunct program at the University of
Ottawa, Ready and Wesche (1992) found that, initially,
adjunct students believed that the lectures were difficult
to understand but the students' comprehension improved
rapidly. The activity rated as the most helpful for content
learning included reviewing class lectures; the activities
that students found the most helpful for language learning
included oral and written production for communicative
purposes (e.g., written summaries of main points, essays
comparing information from different sources).
Valentine & Repath-Martos (1992) investigated the
relevance of the materials and tasks in a simulated adjunct
in which the academic content and reading and writing
requirements of general education courses formed the basis
for course material in the ESL course. The results of a
survey showed that the highest rated academic skills taught
in class included paraphrasing/summarizing, followed by
writing in class essays, and paced and timed readings.
Writing and reading comprehension were perceived as very
important skills by nearly 90% of the students enrolled in
the course, and reading speed was perceived to be very
important or important by nearly 50% of the -students. The
58
students in the program believed that not enough class time
was spent on grammar or vocabulary teaching. * •
Adamson (1990) assessed how well a precourso— modified
adjunct— prepared students for content classes. In his
investigation, Adamson found that, although the term papers
of his precourse students compared well with those of a
control group, the precourse students did not do well on
open-book, open-note examinations, still, the precourse
students were enthusiastic about the experience because they
thought that the course prepared them well for academic-
level work.
A modified version of the adjunct model was found to be
effective in helping students improve their reading ability
(Hudson, 1991). In this version of the model, the adjunct
course was not linked to a specific content course; instead,
its materials were drawn from a variety of courses that
students had taken or were taking. The adjunct course
participants outperformed uninstructed students on three
reading subtests. According to Hudson, these findings
suggest that reading comprehension and general reading
ability can be improved through a content comprehension
approach.
Finally, Duffy (1995) conducted a case study of an ESI.
student enrolled in an adjunct class. In her investigation,
Duffy found that, as a result of the adjunct model, the ESL
59
student gained in linguistic and discourse control, and in
concept development. Although Duffy's results are
preliminary, they are indicative of the effectiveness of the
model.
The previous sections provided a review of the
literature in the area of summary writing, including its
cognitive operations, as well as empirical research in
summary writing in the Ll and L2 reading and writing fields.
The following section presents a number of studies that may
be different in their foci but that are directly related to
the L2 student population that participated in this study.
Previous Studies Related to the Student Population
Represented_J.n_This_Studv
An extensive review of the literature on the reading
and the composing process of L2 students revealed that most
investigations to this date have focused on three student
populations. The first includes international students,
also known as elective bilinguals (Vald6s, 1992). These are
former EFL students who "have elected to enroll" (p. 113) in
American universities, (c.f., Carrell, Pharis, & Liberto,
1989; Connor & McCagg, 1987; Hudson, 1988; Kaplan, 1966;
Reid, 1990; Zimmerman, 1994). The second population
investigated in the literature includes Incipient
60
bilinguals, that is, immigrant students who have acquired a
first language and are in the process of acquiring English
in the U.S. (Valdds, 1992) (c.f., Edelsky, 1982, 1983, 1993;
Hern&ndez, 1991; Padr6n, 1992; Padrdn, Knight, & Waxman,
1986; Padr6n & Waxman, 1988; Urzua, 1987). The third
population that has been researched includes students in FL
contexts (c.f., Bensoussan & Kreindler, 1990; Cohen, 1993;
Sarig, 1987, 1993).
In contrast to these comprehensive lines of research,
very few studies have focused on what Valdds calls American
bilingual minority students (p. 113), that is, students who
can function in two languages, although the level of
proficiency in the two languages may differ for different
individuals. American bilingual minority students differ
from international students in their backgrounds and
characteristics, as well as in their attitudes, academic
experiences, and preparation for the demands of higher
education (Basham, Ray, & Whalley, 1993; Gaies, 1991;
Murray, Nichols, & Heisch, 1992). Although these students
graduate from U.S. high schools and are relatively
proficient in English, they may be thought to need special
instruction to develop their academic writing skills in
English. These immigrant students often do not participate
in mainstream composition classes and are enrolled in
courses designed for culturally different students (Basham,
61
Ray, & Whalley, 1993, p. 302). At other tines, these
students may find themselves in regular composition programs
because the sections for ESL students are full or they do
not see themselves as foreign, but American students
(Hurray, Nichols, & Heisch, 1992)2.
This section presents a review of the research related
to bilingual students in high school- and college-level
programs. Three points should be made about this section.
First, although the review is intended to be extensive, it
is not exhaustive. Second, although different labels are
used by different researchers (i.e., bilingual vs. ESL
students3), the labels may, in fact, describe comparable
student populations. To complicate matters further,
researchers may draw conclusions about the students' degree
of bilingualism based upon the students' perceptions. Third,
the studies summarized in this section differ from one
another in their foci and in the racial or ethnic background
of their student populations. Yet, all of the studies are
linked in that they relate to immigrant, as opposed to
international, students enrolled in high school and college
programs.
In a study at San Jos6 State University, Hurray,
Nichols, and Heisch (1992) administered a survey to
determine whether there was a correlation between self-
reported language use by Asian immigrant and refugee
62
students and their performance on an essay examination. The
results of the study showed that, for beginning-level
students, there was a correlation between using a language
other than English and the students' performance on the
essay examination. For advanced-level students, there was a
correlation between their performance on the essay
examination and their use of a language other than English
when reading books or newspapers, watching television, and
writing business letters.
Four studies focused on L2 writers. In the first
study, Basham, Ray, and Whalley (1993) investigated how
bilingual/bicultural university students represented tasks
to themselves and how they planned for writing. The study
focused on native speakers of Chinese, Alaska Native
students— most of them reporting themselves bilingual— and
Latino students, most of whom were native Spanish speakers.
Basham, Ray, and Whalley found that, while Latino students
oriented themselves toward the topic, Asian and Alaska
Native students placed more value on the text and the task.
The second study, conducted by Montafto-Harmon (1988)
examined the compositions of four different groups: (1)
native Spanish speakers living in Mexico writing in Spanish,
(2) Chicano students— English dominant students— writing in
English, (3) ESL students— Spanish dominant— writing in
English, and (4) monolingual Anglo students writing in
63
English* The ESL students in Montafto-Harmon•s study
reported functioning in Spanish at home and in English for
formal purposes. As was the case for the ESL students, the
Chicano students reported using some Spanish for informal
purposes, although they reported using more English than
Spanish. The Chicano students felt that their skills in
English were stronger than those in Spanish for formal
situations, and tended to function mainly in English,
especially at school.
In her study, Montafto-Harmon found that the writing of
Chicano students differed significantly from the writing of
the other three groups in its conversational and casual
tone. Although Montafto-Harmon found that the compositions
of ESL students contained conversational markers, she
concluded that they were not as casual as the compositions
written by Chicano students. The ESL students' compositions
had digressions, but they were not as frequent in number as
in the Chicano group or among the Spanish writers in Mexico.
As reported in Ferris (1992), Patthey-Chavez (1988) reached
similar conclusions in a study of persuasive compositions
written by Anglo and Chicano high school students.
The last study identified in the literature
investigated how immigrant students from the Pacific Rim and
Mexico placed in ESL composition classes reacted to their
instructors' feedback (Ferris, 1995). Ferris found that
64
students paid more attention to teacher feedback on
^ preliminary drafts of their essays. Students also used a
number of different strategies to respond to their teachers'
comments and had problems in understanding the comments.
According to Ferris, these findings suggest that teachers
should be more direct in responding to their students.
Summarv_of__This_Chapter
This chapter can be summarized into four main areas.
First, research in the fields of LI reading and cognitive
psychology has shown that summary writing is a distinct
reading-to-write task that can be taught. Second, a variety
of studies in the L2 reading and writing fields have
investigated the summarization strategies of L2 students and
have, in general, confirmed LI findings. Third, the effect
of instruction on the summarization strategies of
"underprepared" L2 student populations has not been widely
investigated. Fourth, the L2 reading field has shown that
such variables as reading and cognitive ability, level of L2
proficiency, and text structure can affect the successful
completion of a reading-to-write task.
Notes
1. Another line of research has focused on the notion of
text simplification (e.g., Lautamatti, 1987; Swales, 1990).
Since this body of literature is not directly related to the
present investigation, it is not addressed in this study.
65
2. The tendency of U.S. tertiary institutione to lump
together in classes all persons whose English usage is not
"native academic" is both socially and educationally
impoverished. Some of these students are indeed native
speakers of English but they speak a non-standard variety;
others are genuine bilinguals, speaking two languages
equally fluently, though one or both of the varieties they
speak may be non-standard; some are monolingual speakers of
some language other than English and are learning English as
a second language. The methodologies appropriate to each
group will differ. Equally important, however, is the
sociology of the classroom, because it is unlikely that the
groups share self-perceptions or motivations for language
learning, or that they will collaborate across groups in the
classroom (R. B. Kaplan, personal communication, February 6,
1995).
3. The terms billngual/ESL are confusing because they may
conflate different student populations. Key variables lost
by such terms are: age, gender, prior educational
background, prior second language skills, length of time in
the U.S., etc. For example, a fifteen-year old girl from a
small village in Mexico who received only fifth-grade
education will be very different from a fifteen year old boy
raised in East Los Angeles who has had nine years of
compulsory education, though they may end up in the same
class (R. B. Kaplan, personal communication, February 6,
1995).
66
CHAPTER III
METHODOLOGY
This study was designed to examine the effect of
explicit, adjunct, content-based second language (L2)
instruction on the summarization strategies of
"underprepared" non-native speakers of English (NNS),
university-level native Spanish-speaking (NSS) freshmen
enrolled in three general education courses. Specifically,
the performance of a group of 20 "underprepared" NSS
freshmen in a treatment group (Experimental Group) was
compared with the performance of 20 "underprepared” NSS
freshmen in Control Group 1 and ten "able" NSS subjects in
Control Group 2. The task required the subjects to
summarize a content-specific article, which contained
features of descriptive and comparative texts. The
Experimental Group, Control Group 1, and Control Group 2
were compared for differences in their overall performance
on the summarization task and for differences in the number
of main ideas reproduced. In addition, summary protocols
produced by the subjects in the three groups were analyzed
for differences in the strategies employed to reproduce and
to combine main ideas.
This chapter is organized as follows: (1) the subjects
participating in this study, including their selection, and
their equivalence in content knowledge and in their
67
awareness of the requirements of a summarization task; (2)
the intervention, including the intervention procedures and
the instructional materials; (3) the instrumentation; (4)
the data collection procedures, including the reading and
the summarization tasks; (5) the discourse analytic
procedures, including the identification of idea units and
main ideas, and an explanation of the assessment procedures,
including holistic, summarization efficiency, and analytic
assessment; and (6) the statistical analyses procedures.
Subjects
This study was designed to investigate three groups of
subjects enrolled in one of three university-level general
education courses (Health Science 150, Cultural Anthropology
250, or Animal Biology 155) at California State University,
Los Angeles (CSLA). The first group (Experimental Group)
was composed of 20 "underprepared" subjects, all NSS
freshmen, who were concurrently enrolled in one of the three
general education courses and received eight hours of
explicit summarization-strategy instruction through their
participation in an adjunct, content-based L2 instruction
program entitled Project LEAP: Learning English for
Academic Purposes.
The second group, Control Group 1, was composed of 20
"underprepared" subjects, all NSS freshmen, who were
68
enrolled in one o£ the three general education courses, but
did not receive any supplementary instruction. The third
group, Control Group 2, was composed of ten NSS, nablen
subjects— formerly considered "underprepared"— who
participated in one of the three general education courses.
As was the case with Control Group 1 students, Control Group
2 students did not receive any supplementary instruction.
Table 1 presents demographic information on the subjects of
this study.
69
Iable_l
Demographic Information on the Subjects
Group
Experimental Control 1 Control 2
a
%
n
%
n «
Age
20 and below 17 85 19 95 7 70
21 to 25 3 IS 1 5 3 10
Gander
Male 7 35 12 60 2 20
Female 13 65 8 40 8 80
Class in college
Freshmen 20 100 20 100 2 20
Sophomore 6 60
Junior 2 20
Country of birth
United States 7 35 9 45 6 60
Mexico 9 45 10 50 2 20
B1 Salvador 2 10 1 10
Nicaragua 1 5 1 10
Panama 1 5
Guatemala 1 5
Length of Residence
in the U.S.
2 to 4 2 10 1 5
5 to 7 3 15 3 15
8 to 10 2 10 4 20 2 20
more than 11 6 30 3 15 4 40
since birth 7 35 9 45 4 40
U.S. schooling since
Kindergarten 9 45 10 50 5 SO
Elementary school 6 30 4 20 5 50
Junior high school 1 5 S 25
High school 4 20 1 5
Language programs
ESL 4 20 5 25 2 20
Bilingual 7 35 6 30 4 40
Only English 4 20 5 25 2 20
ESL and bilingual 5 25 4 20 2 20
Note, q ■ total number of subjecte within a given group; % *
percentages within a given group.
Table 2 includes the subjects' self-reported
perceptions of their proficiency in English.
70
Table 2
Subjects* Perceptions of_their_Proficiehcv In English
Language Ski 11/Group Very Well Well Mot Well
n
%
11
«
11
«
speak English
Experimental 9 45 9 45 2 10
Control 1 S 25 12 60 3 15
Control 2 4 40 6 60
Understand spoken English
Experimental 11 55 8 40 1 5
Control 1 9 45 10 50 1 5
Control 2 S SO 5 50
Read in English
Experimental 10 50 7 35 3 15
Control 1 7 35 11 55 2 10
Control 2 5 50 5 50
Write in English
Experimental 7 35 8 40 5 25
Control 1 4 20 11 55 5 25
Control 2 4 40 6 60
Speak English in class
Experimental 7 35 11 55 2 10
Control 1 6 30 11 55 3 15
Control 2 3 30 7 70
Understand class lectures
Experimental 6 30 12 60 2 10
Control 1 6 30 14 70
Control 2 5 50 5 50
Read academic textbooks
Experimental 6 30 10 50 4 20
Control 1 5 25 13 65 2 10
Control 2 4 40 6 60
Complete writing
assignments in English
Experimental 4 20 8 20 a 40
Control 1 1 5 14 70 5 25
Control 2 2 20 7 70 1 10
Note, n “ total number of subjects within a given group; % ■
percentages within a given group.
Table 3 presents the subjects' self-reported
perceptions of their proficiency in Spanish.
71
Table__3
Subjects* Perceptions of their Proficiency in Spanish
Language Skill/Group Very Well Well Not Well
n
t
n
%
n
«
Spaak Spanish
Experimental 10 SO 10 50
Control 1 9 4S 7 35 4 20
Control 2 2 20 8 80
Understand spoken Spanish
Experimental 13 65 7 35
Control 1 9 45 9 45 2 10
Control 2 4 40 5 50 1 10
Read in Spanish
Experimental 9 45 11 55
Control 1 6 30 12 60 2 10
Control 2 1 10 a 80 1 10
Write in Spanish
Experimental 5 25 n 55 4 20
Control 1 6 30 8 40 6 30
Control 2 7 70 3 30
Note, n “ total number of subjects within a given group; % ■
percentages within a given group.
Selection of Subjects
The selection of the subjects who participated in this
study was accomplished in three phases. First, prior to the
Winter 1994 Quarter, records of all of the subjects enrolled
in Health Science 150, Cultural Anthropology 250, and Animal
Biology 155 were obtained. The records included the
subjects' scores on the Scholastic Achievement Test (SAT),
the California State University (CSU) English Placement Test
(EPT), and the Gates-MacGinitie Reading Test— for those
subjects who participated in the Educational Opportunity
72
Program (EOP). Records of the courses that all of the
subjects had taken were also obtained.
The second phase consisted of examining all of the
records, and identifying "underprepared" freshmen and
"able”— formerly "underprepared"— subjects. "Underprepared"
subjects were those who, upon admission to the CSU system,
had not been allowed to enroll in the Freshman English
Composition class and, at the time of this study, were
required to enroll in a minimum of one pre-baccalaureate
reading/writing course, prior to enrolling in the Freshman
English Composition class.
To be required to enroll in pre-baccalaureate
reading/writing courses, subjects had to score:
1. 150 or below on the Reading Skills section of the
EPT, or
2. 350 or below on the Verbal section of the SAT, or
3. 56 or below on the Gates-MacGinitie Reading Test
(if the subject was an EOP participant).
Because the reliability of the test scores was a
concern, consistency of results across the different tests
was a consideration. Specifically, to be included in this
study, the subjects' scores had to fall below the cut-off
point on two of the tests. The reason for looking at two,
instead of three tests, was that in many cases, the
73
subjects1 records did not include scores for all three
tests.
As already indicated, the second phase in the subject
selection process involved the identification of "able"—
formerly considered "underprepared"— subjects. For the
purposes of this study, "able" subjects were those who, upon
their admission to the CSU system, had been required to
enroll in pre-baccalaureate courses, and had, at the time of
this study, fulfilled the Freshman English composition
requirement.
At this point, potential pools of subjects were formed,
one consisting of "underprepared" freshmen, and the other
consisting of "able" subjects. The third and final phase in
the subject selection process involved the identification of
NSS subjects through the administration of a Student Profile
Questionnaire aimed at collecting comprehensive information
about the subjects, including language background (see
Appendix A). While all of the students in the three general
education courses completed the survey, only those students
who reported Spanish as their first language remained in the
final pool. However, as will be observed in Table 4, not
all of the subjects who reported Spanish as their first
language used only Spanish prior to entering school.
74
Table 4
Home Language Background
Group Language Used Before School
Spanish Only Mostly Spanish,
Sometimes English
n % n %
Experimental 18 90 2 10
Control 1 14 70 6 30
Control 2 9 90 1 10
Note, n « total number of subjects within a given group;
% “ percentages within a given group.
Equivalence of the Three Groups_in Content Knowledge
The prior knowledge of the Experimental Group, Control
Group 1, and Control Group 2 was checked to determine
whether it was equivalent, as indicated by the same degree
of prior knowledge of the topics to be used in the
summarization task. The measure of prior knowledge of each
of the topics in the three general education courses was
obtained prior to the completion of the pre-intervention
summarization task. For the purposes of checking the
equivalence among the three groups, the subjects in each of
the three general education courses were asked to write what
they knew about the topic about which they would be reading
(Spivey, 1983)(see Appendix B). Subjects had five minutes
to perform this task, which was scored by awarding 0, 1, or
75
2 points per subject. The subjects that were included in
this study were those who obtained a score of 0, which
indicated that they either gave no response or no correct
information about the topic (Alexander, Kulikowich, &
Schulze, 1994). It is not known whether, in all cases, a
score of 0 was the result of genuine lack of knowledge about
the topic or of the constraints caused by the five-minute
time limit. However, none of the subjects who obtained a
score of 0 complained that they did not have enough time to
complete the task.
Equivalence of the Three Groups in Awareness of the Task
Demands
Prior to engaging in the summarization task, following
Winograd's (1984) procedures, the three groups were
questioned as to what they understood to be a "summary of an
article." To answer this question, the subjects in each of
the three groups were given five minutes to write down their
responses.
In general, the subjects in this study agreed that a
summary should be brief and include the main ideas presented
in an article. Two examples given below from the corpus
illustrate this viewpoint:
To tell about the main points, begining, middle
and end without much detail, yet making it clear
what is all about. (#58)
76
... when you need to write down the main idea of
the article. The important information that
related to the topic of the article. (#04)
In contrast, very few students emphasized the idea of
paraphrasing when writing a summary. This is illustrated in
the example below.
What I understand is that it [a summary] is a
brief explanation of what the article is about, of
course in your own words. (#01)
The Intervention
Procedures
The subjects in the Experimental Group received
explicit instruction in summarization strategies for eight
hours over a period of nine weeks during the Winter Quarter
session of 1994.
Summarization instruction proceeded in six steps.
First, students read a summary of a content-specific article
and, working in groups, evaluated the summary using a
checklist adapted from Casazza (1993) and Kirkland and
Saunders (1991) (see Appendix C). Second, students received
a one-page handout that included strategies for summary
writing (adapted from Brown 6 Day, 1983; Brown, Day, &
Jones, 1983; Hare & Borchardt, 1984; Kirkland & Saunders,
1991) (see Appendix D). Through class discussion, students
became aware of how, when, and why, the summarization
strategies were to be applied (O'Malley & Chamot, 1990;
Rosenshine & Meister, 1994; Shih, 1992; Weinstein & Mayer,
77
1986). Summarization instruction included the following
strategies:
1. Identifying main ideas;
2. Deleting redundant and minor details;
3. Combining ideas in different paragraphs or
sections of the article;
4. Combining lists; and
5. Revising and editing the summary.
Then, employing a content-specific article, and using
the think-aloud technique (Mist & Kirby, 1986), the
instructor modeled summarization-strategy use. Third, a
content-specific article was distributed to the class, and
working in dyads, students wrote a summary, which the class
evaluated using the handout: Eight Steps to Evaluating a
Summary (see Appendix E). Fourth, students received a copy
of another content-specific article and a poorly written
summary based on the article. Using the handout in Appendix
E, and working in dyads, students evaluated the summary
first and, as a follow-up activity, enhanced the summary.
Then, the whole class evaluated the changes suggested by the
groups. Fifth, students self-selected a content-specific
article and, as a homework assignment, students wrote a
summary which was evaluated by a partner in the following
class session using the Eight Steps to Evaluating a Summary
(see Appendix E). The final step consisted of providing
78
students with a handout that demonstrated effective use of
connecting words in summary writing (see Appendix F). Using
the handout and working in dyads, students decided whether
and how the sentences in their summaries could be connected
to produce a cohesive summary.
As can be observed, students learned how to use the
summarization strategies through assisted performance (Tharp
& Gallimore, 1988), meaning that only gradually— and with
the assistance of the instructor or a peer— did the
Experimental group learn summarization-strategy use. In
addition, instruction was influenced by cognitive
instructional models proposed by Anderson (1983), Clark
(1990), Guido and Colwell (1987). In these models, learning
to use strategies requires direct instructional methods and
student-centered practices.
Classroom instruction included the following steps:
1. Giving students clear objectives;
2. Listing the steps for achieving the
objectives;
3. Modeling the strategy to be used;
4. Affording opportunity for guided practice in a
cooperative environment;
5. Encouraging independent practice; and
6. Redemonstrating, or providing correcting
feedback, when necessary.
79
Materials for Summer ization-Stratecrv Training
The materials used for summarization-strategy training
in the adjunct, content-based instruction courses included
authentic reading passages directly related to the general
education course in which the students were enrolled. For
example, for the purposes of instructing students in summary
writing in the Health Science 150 adjunct class, students
used articles from a list of journals (including The Journal
of Health Education) and health-related magazines (including
Health) provided by the content instructor. The same
procedure was followed for Animal Biology 155 and Cultural
Anthropology 250.
Instrumentation
Materials
The materials employed for the purpose of the study
consisted of a student Profile Questionnaire and three
articles (source texts). The Student Profile Questionnaire,
administered during the first week of class, was aimed at
gathering information about the subjects' general
background, language and school background, and language use
(see Appendix A). The Student Profile Questionnaire was a
refined version of the student questionnaire that had been
piloted with a similar student population the previous
quarter. The final version of the Student Profile
80
Questionnaire included the suggestions made by the three
content instructors and the two Project LEAP co-directors.
The three source texts for the pre- and post
intervention summarization tasks included features
characteristic of descriptive and comparative texts (Meyer,
1981) on topics agreed upon with each of the content
instructors. Word counts for the three articles, all of
them from Science News, were 2144 for the Health Science
article, 2217 for the Cultural Anthropology article, and
2328 for the Animal Biology article (see Appendix G).
Ratios of sentence length to word length indicated that all
of the articles were aimed at the college level (Fry, 1977).
Two other measures of the appropriateness of the
materials were obtained. First, each of the content
instructors read the articles and concurred that they were
appropriate for general education students in their
particular classes. After that, six experienced CSLA
reference librarians read the articles and agreed that they
were typical of articles selected by freshman and sophomore
students when completing their general education research
assignments.
81
Data -Collection-Procedures
The subjects who participated in this study completed
the pre- and post-intervention summarization tasks in two
different stages. One will be referred to in this chapter
as the reading task and the other as the summarization task.
The Reading Task
Upon completion of the background knowledge check,
students in each of the general education courses received a
packet which included: (a) one of the three articles,
depending on the content-area class in which the students
were enrolled, (b) instructions on how to read it (see
Appendix H), and (c) the Student Profile Questionnaire.
Content instructors distributed the reading packet in
class and emphasized the importance of following the
instructions in the packet because, during the following
class, students would complete an in-class writing task,
based on the article.
The Summarization Task
The summarization task was piloted during the Fall
Quarter of 1994 with "underprepared" language minority
students enrolled in "developmental" reading and writing
courses at CSLA. Upon the completion of the pilot task,
students were asked questions to determine the level of
82
difficulty and clarity of the task. Based on these
responses, the task was revised for the purposes of the
final study.
Prior to the Winter Quarter of 1994, the three content
instructors agreed that the summarization task which would
be employed for the purposes of this study would constitute
one of their course requirements. The instructors'
rationale for this decision was that normally they required
outside reading and research to enhance classroom discussion
and to augment the information in the textbook. Thus, the
three instructors agreed that the summarization task would
constitute an appropriate class assignment. Therefore, all
of the students registered in the three general education
courses participated in the pre- and post-instruction
summarization tasks, and received 20 points toward their
final course grade. Make-up sessions were held for those
students who were absent on the day the summarization task
was completed, but those students were not included in this
study.
During the class session immediately following the
session when students received the reading packet, the
content instructors set aside 35 minutes for the
summarization task (see Appendix I). A similar procedure
was followed at the end of the quarter, when students
participated in the post-intervention summarization task.
83
Subieats' consent. The subjects' agreement to
participate in the study was requested immediately after the
completion of the post-intervention summarization task. At
the time of the request, prospective study subjects were
informed that their identity would be protected and that
only group results would be reported.
Data Analyses: Discourse.Analvtic Procedures
Identification of Idea Units and Main Ideas
The three source texts were divided into propositions,
with each proposition including one idea (see Appendix J).
The reason for not relying on punctuated sentences, as done
by some LI researchers (Winograd, 1984, for example) was
that more than one idea may be contained within the sentence
boundary.
Ten "expert" readers, operationalized as university
graduates and graduate students, were asked to read the
propositions and rate their level of importance for a
summarization task. Following Garner's (1982) and Johnson's
(1970) procedures, propositions were rated 3, 2, or 1,
depending upon whether or not they contained very important
information that should appear in some form in a text
summary, moderately important information that may or may
not appear in some form in a text summary, and unimportant
84
information that should not appear in any form in a text
summary*
The ratings awarded by the ten "expert" readers were
nearly identical for the important propositions. Within
this study, important propositions were taken to be those
rated 3 by a minimum of nine readers. Those propositions
were then recombined to express main ideas in the source
texts (see Appendix K).
Once the main ideas were reconstructed, they were
deconstructed into idea units (see Appendix L). For the
purposes of this study, an idea unit was defined as "an
isolatable element of a discourse supporting the topic of a
discourse" (Kaplan, personal communication). Categories for
idea units identified in Carrell (1985), Kroll (1977), Johns
(1985), and Johns and Mayes (1990) were adapted for the
purposes of this investigation. What follows are italicized
examples for each of the ten idea unit categories identified
in the corpus.
1. A main clause was counted as one idea unit
including— when present— a direct object:
Ongoing studies of light treatments, biological
clock, and eye's sensitivity make better
understanding ... (#54)
This disease is the number one killer ... (#11)
2. Noun clauses introduced by modality markers were
counted as separate idea units. (As explained by Lautamatti
85
[1987], in this way, the focus remained on the discourse
topic, instead of the topicalized or fronted element of the
sentence.)
It shows that men and women are affected ... (#11)
Host assume that the treatment must somehow alter
circadian rhytms. (#35)
3. Full and reduced, restrictive and non-restrictive,
relative clauses were counted as separate idea units.
...and the Gebusi kill the relative who live with
them. (#30)
. . . , which stands for Seasonal Affective Disorder,
... (#38)
...4 out of 5 Gebusi murders involved someone
branded of sorcery. (#55)
4. Adverbial clauses and phrases were counted as
separate idea units.
..., because people with strong SAD stay
indoors,... (#38)
After having a heart attack,...(#104)
5. Appositives were counted as separate idea units.
A seasonal depression known as SAD-seasonal affective
disorder-where people get depressed by the winter &
when winter nights are the longest. (#53)
6. Phrases that occurred in sentence initial position
followed by a comma or phrases that were set off from the
sentence with commas or dashes were counted as separate
ideas units.
In his study, ... (#30)
86
7. Optional and/or heavy prepositional phrases were
counted as separate idea units (Carrell, 1985j.
...sorcery killings in the Gebusi tribe...(#30)
8. When a verb was followed by an infinitive
construction or gerundive, the second verb phrase was
counted as a separate idea unit.
...Heart disease nay be nore difficult to diagnose
in women* (#01)
...Women wait much longer to see the doctor. (#06)
9. Multiple subjects and multiple objects were counted
as separate idea units.
Food is shared between them and anger and violence
is frown upon. (#65)
10. Transitional phrases in initial sentence position
were counted as separate idea units.
Also, because people with strong SAD stay indoors
very often, now doctors recommend morning outdoor
walks to help SAD patients. (#38)
But their murderous ways cannot be explained. (#55)
Holistic Assessment
A 6-point rating scale for summary protocols was
developed to analyze the holistic performance of the
subjects (see Appendix M). For the rubric used in this
study, many of the features mentioned (e.g., organization,
development) were identified in the scoring guidelines for
the TOEFL Test of Written English (TWE). Features directly
87
related to summary writing were identified in the LI reading
literature, -and incorporated into the rating scale.
Prior to this study, in the Spring Quarter of 1994, two
raters piloted a second version of the Holistic Rating Scale
for Summary Protocols with summaries written by students in
Political Science 150* Interrater reliability for this
pilot test was .93. Based on the results of this pilot
administration, the rating scale was further revised.
Selection and training of raters. Three experienced
English as a second language (ESL) teachers scored the
summary protocols over a six-day period. Prior to engaging
in the analysis of the summary protocols, three norming
sessions of three hours each were held. During the norming
sessions, the raters and the researcher analyzed a large
number of summary protocols not included in the current
study and identified model summary protocols for each of the
six scores (one through six). To eliminate potential bias
deriving from handwriting and neatness, the raters were
given typed copies of the summary protocols written by
subjects.
All summary protocols were rated independently by the
three raters. The holistic scores used in the analysis were
the sum of the three independent scores, which ranged from
three to 18. Because classical theory was not employed for
the purposes of calculating interrater reliability, the
88
section below is devoted to an explanation of the
calculation of reliability.
Reliability'calculation. As explained by Marcoulides
(1989), "classical test theory permits only one [italics
added] source of error in a measurement procedure (either
occasions, raters, or items), and so the combined sources of
error cannot be determined" (p. 116). Specifically, "G
theory recognizes multiple sources of measurement error,
estimates each source separately, and provides a mechanism
for optimizing the reliability" (Shavelson, Webb, & Rowley,
1989, p. 923). Because this study involved three raters and
a test-retest design, reliability was estimated using
Generalizability (G) theory (Cronbach, Gleser, Nanda, &
Rajaratnam, 1972). The relative coefficient in this study
was .83. In G theory, relative coefficient incorporates all
the sources of variation that include subjects (except the
universe score) (Marcoulides, 1989).
Summarization Efficiency Assessment
Two summarization efficiency ratios were calculated to
determine whether the summary protocols produced by the
Experimental Group and by the two Control Groups differed
(1) in the number of main ideas in relationship to the
number of words included in the summary protocols and (2) in
89
the number of main ideas in relationship to the number of
idea units used to express those main ideas.
It was expected that a high mean, as indicated by the
mean ratio of the number of main ideas to the total number
of words in the summary protocols, would be an indicator of
summarization efficiency (Garner, 1982). At the same time,
it was expected that a high mean, as indicated by the mean
ratio of the number of main ideas to the number of idea
units used to express them, would provide an indication of
the writers' ability to express a high number of main ideas
using a small number of idea units.
Garner's (1982) formula of summarization efficiency was
adapted for these calculations. For the purposes of
calculating the mean ratios, an attempt was made to divide
the summary protocols belonging to each of the three groups
into two subgroups, each including a fairly equal number of
protocols. A summary protocol belonged in one group if it
had received a sum-score of 8 points or below on the
Holistic Rating Scale for Summary Protocols. A sum-score of
8 or below indicated that, in the judgment of the raters,
the summary writer did not yet demonstrate competence in
summary writing. By contrast, a summary protocol belonged
in a different group if it had received a sum-score of 9 or
above on the Holistic Rating Scale for Summary Protocols. A
sum-score of 9 or above indicated that, in the judgment of
90
the raters, the summary writer demonstrated some developing
competence in summary writing. This task was undertaken
with the understanding that, for the purposes of comparing
summary protocols with low and high holistic scores,
subgroups with equal numbers were desirable. However, from
the outset, it was understood that the scores' distribution
would be skewed since Control Group 2, consisting of "able"
subjects, was expected to obtain higher holistic scores,
most probably falling on the upper end of the rating scale.
Once the pre- and post-intervention summary protocols
for the Experimental Group and for Control Groups 1 and
Control Group 2 were each divided into two subgroups, the
number of main ideas and the number of words included in the
summaries were counted. Next, the efficiency of
summarization mean ratio was calculated by tabulating the
number of main ideas present in each of the three groups'
summary protocols divided by the number of words included in
the summary protocols of each of the three groups. Then the
second efficiency of summarization mean ratio was calculated
by tabulating the number of main ideas present in each of
the three groups summary protocols divided by the number of
idea units used to express them.
91
Analytic Assessment
The summary protocols were also analyzed to determine
the strategies used by the NSS subjects to reproduce the
main ideas in the source text, prior to and following
summarization instruction. For this purpose, an analytic
grid was developed and categories for analysis were
identified. This grid combined features of:
the grid by Johns and Mayes (1990),
which was based on:
the scale used by Winograd (1984)
which was based on:
Kintsch and van Dijk's (1978)
classification of macrooperations.
In addition, the analytic grid included categories employed
by Campbell (1987) in her study of L2 academic writers.
This part of the study included two units of analyses—
main ideas and idea units contained in the main ideas. Main
ideas were analyzed for two categories: (1) adjacent main
ideas, and (2) conjunctive relationships between two
adjacent main ideas. The idea units contained in the main
ideas were analyzed for three categories: (1) number of
idea units reproduced, (2) strategies used to combine the
idea units contained in the main ideas, and (3) strategies
used to reproduce the idea units. In turn, each category
was divided into detailed subcategories.
The category Adjacent Main Ideas included instances of
two main ideas, one immediately after the other, whether or
not they were separated by a punctuation mark. The category
92
Conjunctive Relationships between Two Adjacent Main Ideas
included adjacent main ideas linked by conjunctions that did
not appear in or were different from those in the original
text (Spivey, 1983). The following list of links from the
corpus, though very limited, covers the types of links
included in the subjects' summary protocols. Italicized
examples from the corpus are given.
1. causal - two main ideas linked by a cause/effect
relationship between the content of the two (Halliday &
Hasan, 1976).
...in getting heart disease. So then women usually
have heart attacks 10 years later in life,... (#02)
2. adversative - two main ideas linked by a "contrary
to expectation" (Halliday & Hasan, 1976, p. 250)
relationship.
... and it is understood it is the number one
killer in men. However, it is also known to be...
(#12)
3. additive - two main ideas linked by adding new to
already given information (Halliday & Hasan, 1976).
Also, many doctors believe that... (#06)
The category titled Strategies Used to Combine Idea
Units included instances of Idea Units Reproduced
Individually and Combinations of Idea Units within and
across Paragraphs. What follows are the idea unit
categories and italicized examples from the corpus.
93
1. an idea unit reproduced by itself.
Having a coronary artery bypass surgery gives a women
greater chance of death than men. (#10)
2. an idea unit reproduced in combination with another
idea unit or with two or more idea units from the same
sentence.
This study dealt with SAD which is a seasonal
affective disorder. (#57)
After menopause, the women are in great danger of
heart disease as men. (#04)
3. an idea unit reproduced in combination with one or
with two or more idea units from different sentences in the
same paragraph.
Heart disease is the number one killer for both man and
women. (#11)
But women suffer heart attacks a decade later in life
and this puts women in danger because they may not
recover as quicky. (#05)
4. an idea unit reproduced in combination with two or
more idea units from the same sentence and from different
sentences in the same paragraph.
Heart disease represents the number-one killer of
female & male in the United states. (#20)
5. an idea unit reproduced in combination with two or
more idea units from other sentences in different
paragraphs.
People who experience seasonally recurring depression
(SAD) have feelings of sadness, anxiety and lethargy.
(#59)
94
6. an idea unit reproduced in combination with two or
more idea units from the same sentence and from different
sentences in different paragraphs.
SAD, which stands for seasonal affective disorder,
is known by symptoms of depression, anxiety,
sadness, drowsiness, and difficulty waking up* (#51)
7. an idea unit reproduced in combination with two or
more idea units from the same sentence and from different
sentences in the same paragraph and in different paragraphs.
The article ”Here Comes the Sun” is about a seasonal
depression, called SAD (seasonal affective disorder),
which appears in late November and ends in April of the
following year, (#49)
The category titled Strategies Used to Reproduce Idea
Units included instances of Copies, Near Copies, Quotations,
and Paraphrases. Examples from the corpus are given for
each of the four subcategories.
1. An idea unit was counted as a copy when all of its
content words were the same as those in the idea unit in the
source text.
The ongoing studies of light treatments, the
biological clocks and the eye's sensitivity to light
herald the dawn of a better understanding of seasonal
depression. (#47)
2. An idea unit was counted as a near copy when all of
its content words but one or two were copied from the
original idea unit or when syntax was rearranged (Campbell,
1990).
95
Studies of light treatments, biological clock £ eye
sensitivity draw better understanding of seasonal
depression. (#54)
3. An idea unit was counted as a quotation when
quotation narks enclosed the original idea unit.
" . . a womens risk of developing heart disease
mounts rapidly,...» (#07)
4. An idea unit was counted as a paraphrase when
syntax was substantially different fron that of the original
idea unit.
... women contract the disease at a later age (for
women typically suffer attacks about a decade later in
life.) (#15)
Selection and trainina_of_raters. Two of the holistic
raters and the researcher, working jointly, scored the
sumnary protocols using the analytic grid. In this way, it
was possible to answer questions and solve disagreements in
conference. Prior to scoring the summary protocols, the
raters practiced using the grid for approximately six hours
in two training sessions.
For this part of the analysis, all of the raters were
asked to read one subject's summary protocol at a time and
to determine which of the main ideas from the source text
were reflected in the subject's summary protocol. Once the
three raters agreed that a main idea, free from distortions,
was reflected in the subject's summary protocol, the idea
units employed to express the main idea were analyzed. The
96
same procedure was followed for the 50 pre- and post-
intervention summary protocols.
Statistical_Analvses
Using the Statistical Package for the Social Sciences
(SPSS) for Windows, an analysis of covariance (ANCOVA) was
conducted to test for differences in holistic scoring among
the Experimental Group and Control Groups 1 and 2 on the
summarization task following explicit strategy training. In
this analysis, the dependent variable was the subjects'
holistic scores on the post-intervention summarization task.
The subjects' scores on the pre-intervention summarization
task and the subjects' length of residence in the United
States were used as covariates. Post-hoc ANCOVA tests were
run between the Experimental Group and Control Group 1,
between the Experimental Group and Control Group 2, and
between Control Group l and Control Group 2.
A second ANCOVA was conducted to test for differences
in the number of main ideas reproduced by the three groups
in the summarization task following instruction. The
dependent variable was the number of main ideas included in
the post-intervention summarization task. The number of
main ideas included in the pre-intervention summary
protocols and the subjects' length of residence in the
United states were used as covariates. Post-hoc ANCOVA
97
tests were run between the Experimental Group and Control
Group 1, between the Experimental Group and Control Group 2,
and between Control Group 1 and Control Group 2.
In the following chapter, the findings of the research
study are presented. Analyses of these findings follow in
Chapter V.
98
CHAPTER IV
FINDINGS
This study examined the effect of explicit, content-
based instruction on the summarization strategies of
"underprepared" non-native speakers of English (NNS),
university-level native Spanish-speaking (NSS) freshmen
enrolled in three adjunct courses. Specifically, the
performance of a group of 20 "underprepared" subjects in a
treatment group was compared with the performance of 20
"underprepared" subjects in Control Group 1 and 10 "able"
subjects in Control Group 2.
Chapter IV presents the results of the statistical
analyses of the holistic and main idea performance by the
subjects in the three groups, as well as the findings of the
summarization efficiency analysis. The chapter also
provides a description of the writing strategies— including
reproduction and combination strategies— employed by the
subjects in Health Science 150, Cultural Anthropology 250,
and Animal Biology 155. The final section of the chapter
presents a summary of the findings.
Holistic Assessment
An analysis of covariance (ANCOVA) was conducted to
test for statistical differences in holistic scores attained
by the three groups in their performance on the
99
summarisation task. The dependent variable was the
subjects' scores on the post-intervention summarisation task
and the covariates were the subjects' scores on the pre
intervention summarisation task and the subjects' length of
residence in the United States.
The means, adjusted means, and standard deviations for
the Experimental Group and for Control Groups 1 and 2 on the
pre- and post-intervention summarization tasks are shown in
Table 5. Recall that the holistic score on the
summarization task was the sum-score, ranging from a low 3
to a high 18, of the ratings given by the three ESL raters.
Table 5
Means and Standard Deviations for the Holistic Scores
Groups Prs-Intervention Task Post-Intsrvsntion Task
I
■
■
1
I
■
■
■
■
■
I
■
1
a
■■■■■■■
£J2
ISISSISBI
^SJ.....
Adjusted M
Expsrimsntal 7.60 2.06 12.05 2.82 12.17
Control 1 7.45 2.16 7.95 2.56 8.13
Control 2 9.30 2.21 8.90 1.60 8.31
Results of the ANCOVA statistically confirmed the group
outcome differences for holistic performance on the
summarization activity, £ (2, 45) ■ 16.058, p» .000. Of the
two covariates— the subjects' scores on the pre-intervention
summarization task and the subjects' length of residence in
the United States— only the subjects' scores on the pre-
100
intervention summarization task was significant (p =.044).
The results of the ANCOVA test are summarized in Table 6.
Table 6
Analvsls_of_Covarlance for theHollstic Scores
Source
of Variation
BBBBBBBBBBBBBB
Covariates
Pre-test
Length
in U.S.
BBBBBBBBBB
sum of
Squares
BBBBBBBBBB
25.621
25.218
.403
DF
2
1
1
Mean
Square
12.810
25.218
.403
BBBBBBBB1
F
2.191
4.312
.069
SBBBBBBj
sig of
F
.124
.044
.794
Main effects
Treatment
187.810
187.810
2
2
93.905
93.905
16.058
16.058
.000
.000
Explained 213.430 4 53.358 9.124 .000
Residual 263.150 45 5.848
Total 476.581 49 9.726
Post-hoc ANCOVAS were run to test for differences
between the Experimental Group and Control Group 1, between
Control Group 1 and Control Group 2, and between the
Experimental Group and Control Group 2. These tests
revealed that the Experimental Group's holistic scores were
significantly higher than the those attained by Control
Group 1, £ (1, 36) = 23.393, p » .000. Of the two
covariates, only the subjects' holistic scores on the pre
intervention task was significant (p =.043). Another post-
hoc ANCOVA revealed that the holistic scores attained by the
Experimental Group were significantly higher than those
attained by Control Group 2, £, (1, 26) = 8.519, p = .007.
101
Analysis, of Main Ideas
Table 7 displays the means, adjusted means, and
standard deviations for the main ideas included in the pre-
and post-intervention summarization tasks. When examining
the table, recall that any discrepancies in determining
whether or not a main idea was present in a summary protocol
were resolved through discussion by the three raters. As
can be observed in Table 7, the only group that improved its
performance, as indicated by an increase in the mean number
of main ideas, was the Experimental Group. In fact, the
mean number of main ideas produced by the Experimental
groups increased by over 100%.
Table 7
Means and Standard Deviations for Main Ideas bv Grouo
Pra-Xntarvantion Task Post
i I
i I
i I
i i
i i
i I
i i
i X i
l a i
i a I
1 H 1
1 1
1 C 1
1 o 1
1 - 4 1
1 - I J 1
1 C 1
i a i
i > i
1 U 1
i a i
1 4 i 1
1 C 1
1 H 1
a i
Groups Main Idaas Main Idaas
M .....
Experimental 1.55 1.99
H fifi Adjusted H
■ ■ ■ ■ ■ S feiiN V aB H B K B H U H IH H B
3.75 2.00 3.85
Control 1 1.60 1.57 1.90 1.45 1.97
Control 2 2.80 1.99 2.70 1.71 2.37
As shown in Table 8, the three groups also differed in
the total number of main ideas produced. Table 8 displays
these differences and includes the percentage of main ideas
produced from the maximum possible. To obtain the
percentage, first, the maximum possible number of main ideas
102
which could have been reproduced was calculated by adding
the number of main ideas for the three articles and by
multiplying that number by the number of subjects in each of
the groups. Then, the percentage for each of the three
groups was obtained by dividing the number of main ideas
reproduced by the maximum possible number of main ideas
which could have been reproduced. For example, the total
number of main ideas for the three articles combined was 50.
When multiplied by 20 (the number of subjects in the
Experimental Group), the result was 1000, the maximum
possible number of main ideas that the Experimental Group
could have reproduced in the summary protocols. Finally,
the total number of main ideas reproduced by the
Experimental Group (31) was divided by the maximum possible
number of main ideas (1000).
Table 8
Total Number and Percentage of Main Ideas Reproduced bv
Group
Pre-Intervention Teak Post-Intervention Taak
Group Total Percentage Total Percentage
Number of Ideas Number of Ideas
of Main from the of Main from the
Zdeaa Maximum Ideaa Maximum
Reproduced Possible Reproduced Possible
■■■■■■■■■■■■■IVIEIHUSIBNIHISIBIHIBNSBHW
Experimental 31 3.10 75 7.50
Control 1 32 3.20 38 3.80
Control 2 28 5.60 27 5.40
103
Table 8 reveals that Control Group 2 did not exhibit
major changes in its performance regarding the total number
of main ideas produced, and Control Group l exhibited a 20%
gain. By contrast, the Experimental Group showed the
greatest gain in main idea production across groups, with an
improvement of over 100%.
An ANCOVA was run to test for significant differences
among the three groups in their performance on the
summarization task, as indicated by the number of main ideas
reproduced. In this case, the dependent variable was the
number of main ideas produced in the post-intervention
summarization task and the covariates were the number of
main ideas reproduced in the pre-intervention summarization
task and the subjects' length of residence in the United
States. Results of the ANCOVA test statistically confirmed
the group outcome differences for number of main ideas, £
(2, 45) » 6.521, pa .003. Neither of the two covariates
were significant. The results of the ANCOVA test are
summarized in Table 9.
104
Table 9
Analysis of, covariance for-the. Number. of Main Ideas
Reproduced
Source sum of DF
of Variation Squares
atnm
Covariates 10.853 2
Main idea
pre-test 10.514 1
Length
in U.S. .339 1
Main effects 37.107 2
Treatment 37.107 2
Explained 47.960 4
Residual 128.040 45
Total 176.000 49
Mean F Sig of
Square
irrnrTTrrn— — —
F
BBBttBBBI
5.426 1.907 .160
10.514 3.695 .061
.339 .119 .732
18.554 6.521 .003
18.554 6.521 .003
11.990 4.214 .006
2.845
3.592
Post-hoc ANCOVAS were conducted to test for significant
differences between the Experimental Group and Control Group
1, between Control Group 1 and Control Group 2, and between
the Experimental Group and Control Group 2. The test
results revealed that the Experimental Group reproduced a
significantly greater number of main ideas than Control
Group 1, £ (1, 36) ■ 11.536, p- .002. The differences
between the Experimental Group and Control Group 2, as
indicated by the number of main ideas reproduced, did not
reach statistical significance.
105
Summarization.Efflciency_Assessment
For this part of the study, the number of main ideas
and the number of words included in the summaries were
counted, and the efficiency of summarization mean ratio, as
indicated by the number of main ideas expressed in relation
to the number of words included in the summaries, was
calculated.
The summary protocols that obtained a sum-score of 8 or
below contained 0 to 3 main ideas and 68 to 396 words. On
the other hand, the summary protocols that received a score
of 9 or above contained 0 to 8 main ideas and 55 to 465
words. Table 10 presents the efficiency of summarization
mean ratio for the three groups for the pre- and post
intervention summarization tasks.
Table 10
Efficiency of Surcmari.2flti9n Mean Ratio? Numfrer_of-Mflin
Ideas to the Number of Words by group
Experimental
8-below 9-above
Control
B-below
1
9-above
Control 2
8-below 9-above
Teat M % %
M % h
%
n
% M %
Pre .0037 60 .0092 40 .0035 50 .0095 50 .0086 20 .0135 80
Poat .0000 10 .0273 90 .0044 45 .0105 55 .0028 20 .0136 80
Note. 8-below * aum-acore of 8 or below; 9-above - aum acore of 9 or
above; H - Hean ratio, number of main ideae to the number of worda in
the aummary protocola; % - Percentage of aummary protocola within a
given group.
As can be observed in Table 10, the pre- and post
intervention summary protocols that received a sum-score of
106
8 or below in the Experimental Group and in Control Group l
were quite similar in their mean ratios. Mean ratios of
.0037 and below were indicative of summaries that contained
a very low number of main ideas and a high number of words.
By contrast, summary protocols with a sum-score of 9 or
above obtained higher mean ratios. For example, a ratio of
.0092 or above was indicative of summary protocols that
included a higher number of main ideas and fewer words than
the summaries with a sum-score of 8 and below. The results
of the pre-intervention summarization task presented in
Table 10 reveal that the group that wrote the most efficient
summaries, as indicated by a high number of main ideas and a
low number of words, was, as expected, Control Group 2 (mean
ratio - .0135). By contrast, the results of the post
intervention summarization task in Table 10 show that the
most efficient summaries were written by the Experimental
Group following treatment (mean ■ .0273), whose performance
nearly tripled.
Table 10 also shows that the holistic performance of
the three groups differed. Specifically, Control Groups 1
and 2 did not exhibit major changes in the number of
subjects who received a sum-score of 8 or below or 9 and
above in the pre- and post-intervention tasks. Only 5% of
the subjects in Control Group 1 moved from the 8-or-below to
the 9-or-above sum-score category; and no changes in
107
holistic ratings were observed for the summary writers in
Control Group 2.
In contrast, the performance of the Experimental Group
improved significantly across summarization tasks. While
60% of the subjects in the Experimental Group received a
sum-score of 8 or below on the pre-intervention
summarization task, only 10% of the subjects in the
Experimental Group remained in the 8-or-below category on
the post-intervention task. In other words, the results of
the post-intervention task show that, of the 20 subjects in
the Experimental Group, only two obtained a score of 8 or
below and the other 18 obtained scores of 9 and above
following intervention.
Writing Strategies Analyses
The main ideas reproduced by the Experimental Group, by
Control Group 1, and by Control Group 2 in the pre- and
post-intervention tasks administered in Health Science 150,
Cultural Anthropology 250, and Animal Biology 155 were
isolated, categorized, and analyzed. The results of these
analyses are presented below.
Health Science 159
Reproduction of main ideas. Nineteen main ideas were
identified in the Health Science source text. If all of the
108
19 main ideas had been reproduced by the nine subjects in
the Experimental Group and in Control Group 1, a maximum
possible number of 171 main ideas would have been reproduced
by each of the two groups. In addition, Control Group 2,
consisting of one subject, could have reproduced a maximum
possible number of 19 main ideas. Table 11 displays the
total number and the percentage of main ideas reproduced by
group.
Table 11
Mflin-Ideag in Health Sgience; Total Number and Percentage
fry Group
Group Pro-Intervention Task Post-Intervention Task
Total
Humber
of Hain
Ideas
Reproduced
Percentage
of Hain
Ideas from
the Maximum
Possible
Total
Number
of Main
Ideas
Reproduced
Percentage
of Main
Ideas from
the Maximum
Possible
Experimental 22 12.87 35 20.47
Control 1 21 12.28 25 14.62
Control 2 4 21.05 4 21.05
As can be observed in Table 11, both the Experimental
Group and Control Group 1 produced a higher percentage of
main ideas in the post- than in the pre-intervention
summarization task. However, the difference in percentage
growth was larger for the Experimental Group (7.60) than for
Control Group 1 (2.34). It should be noted that any results
109
for Control Group 2 should be taken with extreme caution
since Group 2 consisted of one subject only.
The three groups exhibited differences in the total
number of adjacent main ideas produced (Table 12). While
the results of the pre-intervention task show that Control
Group 1 included a higher percentage of adjacent main ideas,
the results are reversed in the post-intervention task.
Table 12
Adjacent Main Ideas in Health Science; Total NVffllfreE-flnd
Percentage bv Group
Group Pro-Intervention Task Poet-Intervention Taek
Total Percentage Total Percentage
Number from the Number from the
of Adjacent Total Number of Adjacent Total Number
Hain Ideas of Hain Ideas Hain Ideas of Hain Ideas
Reproduced Reproduced
Experimental 9 40.90 24 68.57
Control 1 10 47.62 16 64.00
Control 2 1 25.00 1 25.00
Additionally, although the Experimental Group connected
adjacent main ideas via conjunctions more often in the pre
intervention task, it exhibited a reduction in the use of
conjunctions. By contrast, Control Group 1 students
exhibited an increase in conjunctive connections of over
100%. In connecting adjacent main ideas, all of the groups
relied on additive conjunctions heavily. Table 13 displays
the frequencies and percentages of conjunctive relations
110
between adjacent main ideas identified in the summary
protocols.
Table 13
Conjunctive Relations between Two Adjacent Main Ideas in
Health.-Science; Frequency and Percentage bv Group
Experimental Group Control Group 1 Control Group 2
Conjunctiva Pro Poat Pra Post Pra Poat
Relation £ % f % f % f % £ % £ *
Cauaal 3 33.33 3 12.50 1 6.25
Adversative - - 1 4.16 - - 2 20.00
Additive 4 44.44 12 50.00 4 40.00 6 60.00
Total 7 77.77 16 66.66 4 40.00 9 B6.2S
Note. £ ■ frequency of occurrencea within a given group; % _ percentage
of occurrencea within a given group.
Reproduction of idea units contained in the main ideas.
The 19 main ideas identified in the Health Science article-
were categorized into 42 idea units. Table 14 displays the
means and standard deviations for the idea units contained
in the main ideas reproduced.
Table 14
Means. Standard Deviations, and Total Number of Idea Units
Contained in the Main Ideas Reproduced in Health Science
Group Pre-Intervention Task Post-Intervention Task
11 £2
Total No.
of I.U.
11 £2
Total No
of I.U.
Experimental 5.22 5.16 47 7.89 4.26 71
Control 1 3.33 2.12 30 4.00 1.80 36
Control 2 7.00 0.00 6 8.00 0.00 8
Note. Total No. of X.U. - total number of Idea units produced as part
of the main ideas.
1 100.00 1 100.00
1 100.00 1 100.00
111
As shown in Table 14, the means for the idea units
produced in the post-intervention task were practically the
same for fche Experimental Group and for Control Group 2,
although the total number of main ideas reproduced by the
three groups were different. However, the mean number of
idea units produced did not help explain whether or not the
groups produced a lower or a higher number of idea units in
their attempt to express the main ideas. Therefore, the
main idea efficiency ratio, as indicated by the mean ratio
for the number of main ideas reproduced to the number of
idea units produced to express the main ideas, was
calculated (Table 15).
Table 15
Main Idea Efficiency in Health_Sclencei Wean Ratios* Number
of Main Ideas to the Number of Idea Units bv Group
lupariuntil Group Control Group 1 Control Group 2
8-bnlow 9-above 8-below 9-above 0-below 9-above
Teat M % H % 1 1 % 1 1 % 11% 1 1 %
Pre .4117 33 .5000 66 .7777 33 .6666 66 — — .6666 100
Poat — — .4929 100 .6000 50 .7307 50 -- — .5000 100
Note. 8-below ■ aum-acore of 8 or below; 9-above ■ aura acore of 9 or
above; U - Kean ratio, number of main ideaa to the number of idea unite
contained in the main ideaa; % ■ percentage of nummary protocola within
a given group.
The summary protocols that received a sura-score of 8 or
below contained 1 to 3 main ideas and 2 to 8 idea units
employed to express the main ideas. The summary protocols
112
that received a sum-score of 9 or above contained 0 to 8
main ideas and 0 to IS idea units.
Prior to this investigation, it was expected that, when
compared to the summary protocols that received a sum-score
of 8 or below, summary protocols that received a sum-score
of 9 or above, would include a higher ratio of main ideas to
idea units used to express them. This would indicate that
the subjects with higher holistic scores produced fewer idea
units to express a higher number of main ideas. In turn,
using fewer idea units to express a higher ratio of main
ideas would be an indication of efficiency in main idea
production. However, as Table 15 shows, this was not the
case in the Health Science data; in fact, the mean ratio
shows that summary writers with high and low holistic scores
produced a somewhat similar number of idea units to express
the main ideas in the summary protocols.
Table 15 also provides information on the holistic
performance of Health Science subjects. In the pre
intervention task, in both the Experimental Group and
Control Group 1, there were three summary protocols with a
sum-score of 8 or below and six summary protocols with a
sum-score of 9 or above. By contrast, for the post
intervention task, all of the subjects in the Experimental
Group received a score of 9 or above, and in Control Group
113
1, only six subjects received a score of 9 or above and
three received a score of 8 or below.
Strategies used_to_ comblne__the_idea__units. Table 16
displays the frequencies and percentages of idea units
reproduced Individually. Three points should be highlighted
in regard to Table 16. First, Control Group 2 did not
produce any individual idea units either in the pre- or in
the post-intervention task. Second, Control Group 1 reduced
the production of individual idea units by over 100%.
Finally, the Experimental Group did not produce any
individual idea units in the pre-intervention task; by
contrast, nearly 6% of the idea units produced in the post
intervention task consisted of individual idea units.
Table 16
Individual Idea Units in Health_Sciencei FEeqviengy....flnfl
Percentage bv Group
Idea Units Experimental Group Control Group 1 Control Group 2
Reproduced Pre Post Pre Post Pre Post
f % f % f % f % f t f t
Individually
- 4 5.63 6 20.00 3 8.33 -
Note. £ ■ frequency of occurrences within a given group; % ■ percentage
of occurrences within a given group.
Figure 2 presents the frequencies of idea unit
combinations within paragraphs, and Table 17 presents the
frequencies and the percentages of idea unit combinations
within paragraphs tor the Health Science groups.
114
I
experimental group
control group 1
control group 2
Idea Unite
Hi 2, man Mn. pra
I I ? HIM MTV pMt
H 2.0r.Mn p r a
I I S 2 . < t f . h o . p o r t
IHa*. HIMMClpra
p o r t
^■3*. H I . tan p r a
H I . mi. p o r t
Hl'.wiMl
HI Mn p r a
■Kural
H I , mh pod
Figure 2. Health Science: Frequencies of idea unit
combinations within paragraphs by group.
US
Table 17
IdeaUnit combinations within_Paragraphs_in_Health_science;
Frequency and Percentage bv Group
Xdaa Unit Experimental Group Control Group 1 Control Group 2
Combination Pra Poot Pro Poat Pro Poat
Sfcratogioa f % f % f % f % f % f %
2, aamo aon. 15 31.91 19 26.76 8 26.67 10 27.77
- - 2 25.00
2, dif. aan. 2 4.26 3 4.22 2 6.67 3 8.33
- - -
-
3+, same aan. 19 40.43 19 26.76 8 26.67 7 19.44 - - 3 37.50
3+, dif. aan. - - S 7.04 2 6.67 1 2.78
- - - -
3+, aamo &
dif* aan. 5 10.64 12 16.90 3 10.00 9 25.00 3 50.00 3 37.50
Total 41 87.24 58 81.68 23 76.68 30 83.32 3 50.00 8 100.00
Moto. f - froquoncy of oecurroncaa within a group; % ■ percentage of
occurroncoa within a givon group; 2, aamo aon. ■ combinationa of 2 idoa
unita from tha aamo aontonco; 2, dif. aon. * * combinationa of 2 idoa
unite from different Bontencaa; 3+, aamo aon. ■ combinationa of 3 or
more idoa unita from tho aamo aontonco; 3+, dif. aon. * * combinationa of
3 or mora idoa unita from different sentences; 34-, same & dif. aon. ■
combinationa of 3 or more idoa unita from tho aamo and from different
aontoncoa.
Four patterns of combinations within paragraphs can be
identified in Table 17. First, all of the idea units
produced by Control Group 2 in the post-intervention
summarization task consisted of combinations within
paragraphs. Second, the Experimental Group reduced the
percentage of combinations of idea units within the same
sentence (two idea units ■ -5.15; three or more idea units -
-13.68) and gained in combinations across sentences within
paragraphs (three or more idea units = 7.04; three or more
idea units from the same and from different sentences ■
6.26). Third, the pattern of combinations within paragraphs
for Control Group 1 was random, it exhibited a reduction in
combinations of three idea units from the same sentence (-
7.23), but it did not show major changes in the percentage
of combinations of two idea units from the same sentence
(1.1 gain). By contrast, the percentage of combinations of
three idea units from the same sentence and from different
sentences for Control Group 1 increased by over 100% (from
10.00% to 25.00%).
Figure 3 displays the frequencies of idea unit
combinations across paragraphs and Table 18 displays the
frequencies and percentages of idea unit combinations across
paragraphs for the Health Science subjects. As reported in
Table 18, the instances of combinations of idea units across
paragraphs were very limited for all groups in both the pre-
and in the post-intervention summarization tasks. However,
the Experimental Group was the most consistent in combining
three or more idea units from the same sentence and from
different sentences in different paragraphs across tasks
(10.63% in the pre- vs. 9.85% in the post-intervention
summarization task). By contrast, for Control Group 2,
there was a drastic reduction in the production of three or
more idea unit combinations from the same sentence and from
different sentences in different paragraphs (16.6% in the
pre- vs. 0% in the poBt-intervention task).
117
J D
experimental group
1
control group 1
control group 2
Idea Units
Hid*, < M . tan p r a
L_j3«. d f . tanpott
^■3*. uim a at mr.
mar pm. p r a
•am* t a t . tan
mar. p a r a , p o a t
| 3 » , tamat at aan
tamatar. p a r a p r a
| 3 * . aamatar. aan
unwMrf. p a r a p o a t
Figure 3. Health Science: Frequencies of idea unit
combinations across paragraphs by group.
Table 18
Idea Unit Combinations across Paragraphs in,Health_Science:
Freguencv and Percentage bv Group
Idea Unit Experimental Group Control 1 Group 1 Control Oroup 2
Combination Pra Post Pra Post Pra Post
Stratagias f t f t f % f % f t f t
3+, dif. san. - - 1 2.70 - -
3+, sama f i dif.
san. in dif para. 5 10.63 7 9.85 - - - 1 16.60 -
3+, sama & dif. san.
in sama & dif. para. 1,2.12- - - - - - - - - -
Total 6 12.75 7 9.85 - 1 2.70 1 16.60 -
Note. f ■ fraquancy of occurrancas within a givan group; t “ parcantaga
of occurrancas within a given group; 3+, dif. san. ■ combinations of 3
idea units from different sentences; 3+, sama f i dif. san'. in dif. para ■
combinations of 3 or more idea units from the same sentence and from
different sentences in different paragraphs; 3+, sama £ dif. sen. in
sama G dif. para. - combinations of 3 or more idea units from the same
sentence and from different sentences in the same and in different
paragraphs.
118
Ideaunit reproduction strategies. Figure 4 displays
the idea unit reproduction strategies, which included
copying, nearly copying, quoting, and paraphrasing. Table
19 presents the frequencies and percentages of idea units
reproduced by copying, nearly copying, quoting, and
paraphrasing.
45
experimental group control group 1 control group 2
■ Copies pra
I I Copies post
■ N e a r copies pre
■ N e a r copies post
■Q uotations pre
DilQ] Quotations post
■ P araphrases pre
(^Paraphrases post
Figure 4* Health Science: Frequencies of idea unit
reproduction strategies by group.
119
Table 19
Idea Unit Reproduction Strateqies_in_Health_Science:
Frequency and Percentage bv Group
Reproduction
Strategies
Experimental Group
Pre Post
f * f %
Control
Pre
f %
Group 1
Post
f %
Control Group 2
Pre PoBt
f * f %
Copies 4 8.51 2 2.81 4 13.33 9 25.00 1 16.66 1 12.50
Near copies 11 23.40 10 14.08 10 33.33 9 25.00 1 16.66 1 12.50
Quotations 13 27.66
- - - - - -
Paraphrases 19 40.43 59 83.09 16 53.33 18 50.00 4 66.66 6 75.00
Total 47 100.00 71 99.98 30 99.99 36 100.00 6 99.98 8 100.00
Hota. f * frequency of occurrences within a given group; % - percentage
of occurrencea within a given group.
As revealed In Table 19, the Experimental Group copied
less frequently than Control Group 1 and Control Group 2
both in the pre- and in the post-intervention summarization
tasks. More importantly, the percentage for copying in the
post-intervention summarization task decreased for both the
Experimental Group (8.51% in the pre- vs. 2.81% in the post
intervention summarization task) and Control Group 2 (16.66%
in the pre- vs. 12.50% in the post-intervention
summarization task). By contrast, the results of the post
intervention summarization task show that Control Group 1
increased the number of copies (13.33% in the pre- vs.
25.00% in the post-intervention summarization task).
Table 19 also shows that the three groups reduced
number of near copies, with the Experimental Group showing
the most drastic reduction (Experimental Group » -9.32,
Control Group 1 = -8.33, Control Group 2 «* 4.16).
120
In the summaries produced by Control Groups l and 2, no
instances of quotations were identified either for the pre-
or post-intervention summarization tasks. By contrast, the
Experimental Group quoted nearly 28% of the information in
the pre-intervention summarization task; however, no
quotations were identified in the post-intervention
summaries the Experimental Group.
Nearly 41% of the idea units produced by the
Experimental Group in the pre-intervention summarization
task were paraphrased. This percentage increased by over
100% in the post-intervention summarization task.
Conversely, the percentage for paraphrasing for Control
Group 1 remained almost the same. Although Control Group 2
paraphrased more often than did the Experimental Group in
the pre-intervention summarization task, the results of the
post-intervention task show that Control Group 2 paraphrased
less often than did the Experimental Group (75.00% for
Control Group 2 vs. 89.03% for the Experimental Group).
Cultural Anthropology 250
Reproduction of main ideas. The Cultural Anthropology
article contained 17 main ideas. If the 5 subjects in the
Experimental Group and in Control Group 1 had reproduced all
of the ideas, a maximum possible number of 85 main ideas
would have been reproduced. In addition, the two subjects
121
in Control Group 2 could have reproduced a maximum possible
number of 34 main ideas. Table 20 displays the total number
and the percentage of main ideas reproduced. However, much
as in Health Science, any findings regarding Control Group 2
should be taken with caution since the Group included only 2
subjects.
Table 20
Main Ideas in Cultural ,Anthropology:__Total Number and
EgEfignSflge. by.,-group
Group Pre-Intervention Task Poat-Intervention Taak
Total
Number
of Main
Ideaa
Reproduced
Percentage
of Main
Ideas from
the Maximum
Poaaible
Total
Number
of Main
Ideae
Reproduced
Percentage
of Main
Ideaa from
the Maximum
Poaaible
Experimental 1 1.18 18 21.18
Control 1 0
o
o
a
S 5.88
Control 2 6 17.65 5 14.71
Table 20 shows that both the Experimental Group and
Control Group 1 reproduced a higher percentage of main ideas
in the post-intervention summarization task than in the pre
intervention summarization task. However, when compared to
Control Group 1, the difference in percentage growth is
larger for the Experimental Group than for Control Group 1.
From all of the main ideas reproduced by all of the
three groups in the pre-intervention summarization task, not
one was adjacent to another main idea. By contrast, the
122
results of the post-intervention task reveal that the
Experimental Group was the only group that produced any
adjacent main ideas (seven adjacent main ideas, 38.88%). In
addition, four (57.14%) of those main ideas expressed an
adversative relation, another two (28.57%) showed an
additive relation, and one (14.28%) showed a temporal
relation.
Reproductlon_of—idea__units contained in the main ideas.
The 17 main ideas in the Cultural Anthropology article were
categorized into 41 idea units. Table 21 displays the means
and standard deviations for the idea units contained in the
main ideas produced by the three groups.
Table 21
MeanB. Standard Deviations, and Total Number of Idea Units
Contained in the Main Ideas Reproduced in Cultural
Anthropology
Group Pra-Zntervantion Task Poat-Zntervention Taak
u fifi
Total No.
of Z.U.
h £&
Total
of Z.l
Experimental .80 1.30 4 7.80 7.23 39
Control 1
o
o
•
0.00 0 1.40 1.95 7
Control 2 5.50 .71 11 4.50 3.54 9
■■■«■■«■■■■■■■ ■ ■ ■ ■ ■ M B ■ ■ ■ s . . . . . . . . . . . . . . . . . . . . . . . . . . . . ■ ■ ■ ■ ■ 1■ ■ ■ ■ ■ ■ 1
Note. Total No. of Z.U. ■ total number of Idea unita produced aa part
of the main ideaa.
Both the mean and the total number of idea units for
the Experimental Group increased and were higher than those
for Control Groups 1 and 2 (Table 21). However, as noted
123
for the Health Science corpus, this information was not
helpful in explaining whether or not, in Cultural
Anthropology, there was a relationship between the number of
main ideas reproduced and the number of idea units used to
express them. Table 22 presents the mean ratio for the
number of main ideas to the number of idea units used to
express the main ideas.
Table 22
Main Idea Efficiency in Cultural Anthropology; Mean_Ratlos,
Number of Main Ideas to the Number of Idea Units bv Group
Experimental Group Control Group 1 Control Group 2
8-balow 9-abova 8-balow 9-abova 8-balow 9-above
Teat h % M ' M * M *
Pre .2500 100 — -- .0009 100 — — .0075 50 .0222 50
Poat .0000 20 .0569 80 .0024 80 .0121 20 .0055 50 .0086 50
Note. 8-below " aum-acore of 8 or below; 9-above ■ aum acore of 9 or
above; ft ■ Mean ratio, number of main ideaa to the number of idea unita
contained in the main ideaa; % ■ percentage of summary protocola within
a given group.
Summary protocols with a sum-score of 9 or above
included a range of 1 to 6 main ideas and 2 to 19 idea
units. The summary protocols that received a sum-score of 8
or below were characterized by the inclusion of 0 to 3 main
ideas and the idea units used to express the main ideas
ranged from 0 to 5.
Just as was the case for the Health Science corpus, the
calculation of the main idea efficiency ratio did not render
information that could help differentiate summary protocols
124
with high and low holistic scores. As shown in Table 22,
the results were random at best, with summary protocols with
a sum-score of 8 and below obtaining higher mean ratios than
summary protocols with a sum-score of 9 and above (e.g.,
.2500 vs. .0121).
However, summary protocols produced by the subjects in
the Experimental Group and in and Control Groups 1 and 2
differed in the holistic scores received. Specifically, the
pre-intervention task results indicate that all of the
summary protocols in the Experimental Group and in Control
Group 1 received a sum-score of 8 or below; in addition, one
of the summary protocols in Control Group 2 received a sum-
score of 8 or below and the other a sum-score of 9 or above.
By contrast, the results of the post-intervention
summarization task show that 80% of the summary protocols in
the Experimental Group obtained a sum-score of 9 or above,
while only 20% of the summaries produced by Control Group 1
in the post-intervention summarization task received a sum-
score of 9 or above.
Strategies used to combine idea units. The idea unit
combination strategies employed by the three groups were
analyzed. Control Group 1 was the only group to produce
individual idea units (9.09%) in the pre-intervention
summarization task. By contrast, the Experimental Group was
125
the only group to produce individual idea units in the post
intervention task (7.69%).
In addition, as shown in Figure 5 and Table 23, the
frequencies and percentages of combinations of idea units
within paragraphs for cultural Anthropology were analyzed.
17
16
is
14
13
12
11
10
9
8
7
I6
S
• i
experimental group control group 2
control group 1
Idea Unite
l l , aama Mn pea
1 1 3. i m mo. port
^ l l d f . a a n . pro
[M l 2.01 tan port
I k aama aan. pro
l= j3 * . aoma aan. port
B s + .a a m a l
01. aan. pra
(M is*'aam at
aan. port
Figure 5. Cultural Anthropology: Frequencies of idea units
combinations within paragraphs by group.
126
Table 23
Idea Unit Combinations within Paragraphs inCultural
Anthropology: Frequency and Percentage bv Group
Idas Unit
combination
Stratsglas
Kxpsrimsntal Group
Pra Post
f % f %
Control Group 1
Pro Post
f % f %
Control Group 2
Pro Post
f t f %
2, sains sen. 1 25.00 12 30.77
- 3 4.29 6 54.55 8 8.88
2, dif. sen.
- 4 10.26
- - - 1 11.11
3+, sama sen. 3 75.00 3 7.69
3+, dif. sen.
3+, same &
dif. aan. 16 41.03
• —
3 4.29 4 36.36
_ _
Total 4 100.00 35 89.75
-
6 8.58 10 90.91 9 19.99
Notea, f - frequency of occurrencea within a given group; % ■ percentage
of occurrences within a given group; 2, same sen. • * combinations of 2
idea units from the same sentence; 2 , dif. sen. ■ combinations of 2 idea
units from different sentences; 3 * , same sen. - combinationa of 3 or
more idea units from the same sentence; 3+, dif. sen. ■ combinationa of
3 or more idea units from different sentences; 3+, same 6 dif. sen. “
combinations of 3 or more idea units from the same and from different
sentences*
As can be observed in Table 23, all of the four idea
units produced by the Experimental Group in the pre
intervention summarization task consisted of combinations
within sentences. By contrast, nearly 37% of the idea units
produced by Control Group 2 consisted of combinations within
the same sentence and across sentences in the same
paragraph. However, in the post-intervention summary
protocols, nearly 42% of the idea units produced by the
Experimental Group consisted of combinations within and
across sentences in the same paragraph. Only 4.29% of the
idea units produced by Control Group 1 consisted of
combinations within and across sentences in the same
127
paragraph* Control Group 2 did not produce any combinations
of idea units within and across sentences in the same
paragraph. Neither the Experimental Group, nor Control
Groups 1 or 2 produced any combinations of idea units across
paragraphs.
Idea unit reproduction strategies. Figure 6 displays
the idea unit reproduction strategies, which included
copying, nearly copying, quoting, and paraphrasing. In
addition, Table 24 includes the frequencies and percentages
for the four idea unit reproduction strategies.
30'
25
20
15
$ 1. 0
H Copies pre
I I Copies post
■ I Near copies pre
■ N e a r copies post
■ Q uotations pre
(QQIlQuotetions post
■ P arap h rases pre
(^Paraphrases post
experimental group control group 2
control group 1
Figure 6. Cultural Anthropology: Frequencies of idea unit
reproduction strategies by group.
128
Table 24
Idea_Unit_Repgoduction Strategies in Cultural Anthropology:
Frequency and Percentage bv_Group
Reproduction
Strategies
Bsperiaental Oroup
Pre Post
f % f %
Control Oroup 1
Pre Post
f % f %
Control
Pre
f «
Oroup 2
Post
f %
Copies 2 SO.00 4 10.26 4 57.14
- - 2 22.22
Near copies 1 25.00 7 17.94 1 14.29 2 18.18 3 33.33
Quotations
-
2 5.13
- - - - -
Paraphrases 1 25.00 26 66.66 2 28.57 9 81.81 4 44.44
Total 4 100.00 39 99.99 7 100.00 11 99.99 9 99.99
Mote, f - frequency of occurrences within a given group; % - percentage
of occurrences within a given group.
Three observations must be made regarding the
reproduction strategies in Cultural Anthropology, as
indicated in Table 24. First, both Control Group 1 and
Control Group 2 did not produce any copies in the pre-
intervention task— in fact Control Group 1 did not produce
any idea units in the pre-intervention task. By contrast,
nearly 23% of the information produced by Control Group 2
and nearly 58% of the information produced by Control Group
1 in the post-intervention summarization task consisted of
copies. When compared with both Control Groups, the
Experimental Group reduced the percentage of copies from 50%
in the pre-intervention task to 10.26% in the post
intervention task.
Second, a similar pattern of near copies can be
observed for the three groups. Specifically, the
Experimental Group produced the highest percentage of near
129
copies in the pre-intervention task, and produced fewer near
copies in the post-intervention task. By contrast, both
Control Groups produced near copies more often in the post-
than in the pre-intervention task.
Third, as Table 24 indicates, in the pre-intervention
summarization task, the amount of paraphrasing done by
Control Group 2 was over three times higher than that done
by the Experimental Group (81.81% for Control Group 2 and
25.00% for the Experimental Group). The results of the
post-intervention task show that the Experimental Group
paraphrased more often than Control Groups l and 2. In
fact, nearly 67% of the idea units produced by the
Experimental Group were paraphrased.
Animal Biology 155
RSPEg.dWff.tion included 14 main ideas. If all of the 14 main ideas had
been reproduced by the six subjects in the Experimental
Group and in Control Group 1, a maximum possible number of
84 main ideas would have been produced by each group. In
addition, Control Group 2, consisting of seven subjects,
could have reproduced a maximum possible number of 98 main
ideas. Table 25 displays the total number and the
percentage of main ideas reproduced by group.
130
t
Table 25
Main Ideas in Animal Biology: Total Number and Percentage
bv Group
Group Pra-Intervention Taek Poet-Intervention Taek
Total Percentage Total Percentage
Humber of Hain Number of Main
of Main Ideas from of Main Xdeae from
Ideas the Maximum Ideae the Maximum
Reproduced Possible Reproduced Poaaible
Experimental 8 9.52 22 26.19
Control 1 11 13.09 8 9.52
Control 2 18 18.36 18 18.36
As indicated in Table 25, the total number of main
ideas reproduced by Control Group 2 across tasks remained
the same (18); and the number of main ideas reproduced by
Control Group 1 decreased (11 in the pre- vs. 8 in the post
intervention summarization task). The Experimental Group
was the only group that exhibited an increase in the number
of main ideas reproduced in the post-intervention
summarization task (8 in the pre- and 22 in the post
intervention summarization task).
In addition, much like the groups in Health Science and
Cultural Anthropology, the three groups in Animal Biology
exhibited differences in the strategies used to present
information, as indicated by the number of adjacent main
ideas. Table 26 presents the total number and the
131
percentages of adjacent main ideas produced in the pre- and
^ post-intervention experimental tasks.
Table 26
Adjacent Main Ideas in AnimalBlologv:__TPercentage bv Group
Group Pra-Znterventlon Task Poat-Intorvantion Task
Total Percentage Total Percentage
Number from the Number from the
of Adjacent Total Number of Adjacent Total Number
Hain Ideaa of Hain Ideaa Hain Ideaa of Hain Ideaa
Reproduced Reproduced
Experimental 0 .00 6 27.27
Control 1 1 9.09 1 9.09
Control 2 0 .00 8 4.44
As can be observed in Table 26, the performance of the
Experimental Group in the pre-intervention summarization
task was comparable to that of Control Group 2. Neither
group produced any adjacent main ideas, but nearly 10% of
the main ideas produced by control Group 1 consisted of
adjacent main ideas. In the post-intervention summarization
task, no changes were observed for Control Group 1, and
nearly 5% of the main ideas produced by Control Group 2 were
adjacent to other main ideas. By contrast, nearly 28% of
the main ideas produced by the Experimental Group were
placed immediately after other main ideas.
The three groups also differed in the strategies
employed to connect two adjacent main ideas. Table 27
132
displays the frequencies and the percentages of the
different types of conjunctive links employed by the groups.
Table 27
Conjunctive Relations between Two Adjacent Main Ideas in
Animal Biology; Frequency and Percentage bv Group
Conjunctiva
Relation
Experimental Oroup
Pre Post
f % f %
Control Oroup 1
Pre Post
f \ f \
Control Oroup 2
Pre Post
f « f %
Causal
Adversative
- 2 33.33
Additive
-
2 33.33 1 100.00 2 25.00
Total■
-
4 66 66 1 100.00 4 50.00
Note. £ ■ frequency of occurrancas within a given group; % " percentage
of occurrencea within a given group.
Table 27 shows that the performance of the three groups
in the pre-intervention task was comparable; only Control
Group 1 produced any instances of adjacent main ideas (one
instance), and no conjunctive relation was identified
between the ideas. Post-intervention results reveal,
however, that the Experimental Group connected adjacent
ideas via conjunctions more often than did Control Group 2
(66.66% for the Experimental Group and 50.00% for Control
Group 2). Fifty per cent of the adjacent main ideas
produced by Control Group 2 showed some kind of conjunctive
relation, half of them causal and the other half additive.
The Experimental Group produced four adjacent main ideas—
two linked by an adversative conjunction and the other two
linked by an additive conjunction. By contrast, Control
133
Group 1 produced only one instance of adjacent main ideas,
and the two ideas were linked by an additive conjunction.
Reproduction of idea units contained in the main ideas.
The 14 main ideas identified in the Animal Biology article
were categorized into 30 idea units. Table 28 displays the
means, standard deviations, and the total number of idea
units contained in the main ideas reproduced by the groups.
Table 28
Means, standard Deviations,_and_Total_Number of Idea Units
Contained in the Main Ideas Reproduced_in_Animal Biology
Group Pre-Intervention Task Post-Xntervention Task
h
Total No.
of X.U.
h
Total No
of X.U.
Experimental 2.00 1.90 12 5.83 3.31 35
Control 1 3.00 3.69 18 1.67 1.63 10
Control 2 4.29 3.68 30 3.29 2.69 23
Nots. Total No of X.U. - total number of idea units contained in the
main ideas reproduced*
As was the case for the Health Science and Cultural
Anthropology corpora, the mean number of idea units in
Animal Biology did not provide useful information as to the
efficiency displayed by the subjects in reproducing the main
ideas. The calculation of the main idea efficiency ratio in
Animal Biology was equally inefficient in providing useful
data regarding the relation of the number of main ideas to
the number of idea units used to express them. As shown in
Table 29, the main idea efficiency ratio resulted in random
134
numbers for summary protocols with low and high holistic
scores (e.g., .5000 for summaries with a score of 9 and
above and .7727 for summaries with a score of 8 or below).
It should be noted that, as was the case for Health
Science and Cultural Anthropology, the Experimental Group's
holistic performance on the pre- and post-intervention task
differed from that of the two Control Groups. Specifically,
while most of the Experimental subjects obtained a holistic
score of 8 or below on the pre-intervention task, their
holistic performance improved to the point that most
Experimental subjects obtained a score of 9 and above on the
post-intervention task (Table 29). By contrast, the
performance of Control Group 1 subjects remained constant
across tasks, and the performance of Control Group 2
subjects declined.
Table 29
Main Idea Efficiency in Animal Biology; Mean Ratios. Number
of Main Ideas to the Number of Idea Units bv Group
Experimental Group Control Group 1 Control Group 2
8-below 9-above 8-below 9-above 8-below 9-above
Teat M * M % M ' M ' M « H %
Pre .6000 66 1.0000 33 .0000 33 .6111 66 .5000 14 .6071 86
Poet .0000 17 .6285 83 .0000 33 .8000 66 1.0000 29 .7727 71
Note. 8-below - sum-score of 8 or below; 9-above - eum score of 9 or
above; M - Mean ratio, number of main ideas to the number of idea units
contained in the main ideas; % ■ Percentage of summary protocols within
a given group.
135
Summary protocols with a sum-score of 8 and below were
characterized by the inclusion of 0 to 2 main ideas and 0 to
4 idea units. Summary protocols with a sum-score of 9 and
above contained 0 to 7 main ideas and 0 to 10 idea units to
express them.
Strategies used to combine idea units. Table 30
displays the frequencies and percentages of idea units
produced individually. As can be observed in Table 30, both
the Experimental Group and control Group 2 produced a
slightly higher percentage of individual idea units in the
post- than in the pre-intervention summarization task.
Control Group 1, however, reduced the percentage of idea
units produced individually.
Table 30
individual Idea Units in Animal Biology? FrequenCY-fijld
Percentage_bv_Group
Idea Units Experimental Group Control Oroup 1 Control Oroup 2
Reproduced Pre Post Pre Post Pre Post
f % f % f % f % f t f t
Individually 1 8.33 4 11.42 4 22.22 1 10.00 S 16.66 4 17.39
Note, f ■ frequency of occurrences within a qiven group; % ■ percentage
of occurrences within a given group.
Figure 7 presents two and three idea unit combinations
within paragraphs and Table 31 presents the frequencies and
percentages of combinations within paragraphs for the Animal
Biology summaries.
136
1
I
Idea Unite
^ | 2 . M in a aan. pra
C k u m . a a n poat
His, a t . a an pra
{ H i 2, a t a an poat
H a * , aam a aan. pra
l = i l » a a m a a a n poat
■ H s * . t f aan . pra
I H s * . < M f. a an poet
aam a t
« r . aan. p ra
[M D j *, a a m a *
experimental group control group 2
control group 1
Figure 7. Animal Biology: Frequencies of idea unit
combinations within paragraphs by group.
137
Table 31
Idea Unit Combinations within Paragraphs In Animal Biology.:
Frequency and Percentage bv Group
Idea Unit Experimental Oroup Control Oroup 1 control Oroup 3
Combination Pre Post Pre Post Pre Poat
Strategies f % f % f « f % f % f %
2, same sen. 4 33.33 12 34.28 8 44.44 3 33.33 10 33.33 4 17.39
2, dif. sen. - - - -
1 5.55
- - 3 10.00 2 8.69
3+, same sen* 3 25.00 6 17.14
- - - -
4 13.33 2 8.69
3+, dif. sen.
- -
1 2.86
3+, same 6
dif. sen. 2 16.66 2 5.71 2 11.11
e e _
4 13.33 3 13.04
Total 9 74.99 21 59.99 11 61.10 3 33.33 21 69.99 11 47.81
Note, f • frequency of occurrences within a given group; % « percentage
of occurrences within a given group; 2, same sen. ■ combinations of 2
idea units from the same sentence; 2, dlf. sen. ■ combinations of 2 idea
units from different sentences; 3+, same sen. « combinations of 3 or
more idea units from the same sentence; 3+, dif. sen. ■ combinations of
3 or more idea units from different sentences; 3+, same & dif. sen. ■
combinations of 3 or more idea units from the same and from different
sentences.
Three points must be noted about Table 31. First, the
results of pre-intervention task show that the Experimental
Group and Control Group 2 produced a large number of
combinations within paragraphs. However, the combinations
were mainly done within sentences (over 50% for the
Experimental Group and nearly 47% for Control Group 2).
Second, the results of the pre-intervention task show
that none of the three groups produced any combinations of
three or more idea units from different sentences. However,
in the post-intervention task, nearly 3% of the idea units
produced by the Experimental Group consisted of combinations
from different sentences.
138
Third, the results of the pre-intervention task show
that the Experimental Group produced the highest percentage
of combinations of three or more idea units from the same
sentence and from different sentences in the same paragraph
(16.66% vs. Control Group 1 ** 11.11%, and Control Group 2 «
13.33%). However, the post-intervention task results
indicate that the Experimental Group and Control Group 1
reduced the number of combinations from the same sentence
and from different sentences in the same paragraph, although
the reduction was more drastic for Control Group 1. The
pattern of combinations from the same sentence and from
different sentences in the paragraph for Control Group 2
remained constant.
As can be observed in Figure 8 and Table 32, the
subjects in the three groups combined ideas across
paragraphs.
139
Frggueng'es
Idea Units
^■2, d T . Mn. pm
f It d T . M n. port
TOJs*. d T . M n. port
^|3*. turn A d T . Mn.
i n dr pm pm
3*. m o w A d T . M n.
In d T . p m . port
3 * . M m * A d r . M n.
Mnwur. pm p o r t
axporimental group control group 2
control group 1
Figure a . Animal Biology: Frequencies of combinations
across paragraphs by group.
140
Table 32
IdeaUnit Combinations across Paragraphs In Animal Biology;
Frequency and Percentage bv Group
Idas Unit
Combination
Stratagiaa
Bxpariaantal Oroup Control Oroup 1 Control Oroup 2
Pro Post Pre Post Pra Poat
f % f t f % f % f % f %
2, dif. aan. 1 5.55 1 10.00 1 3.33 2 8.69
3+, dif. aan 1 10.00 2 8.69
3+, aama 6 dif.
aan. in dif. para. 2 16.66 6 17.14 2 11.11 3 30.00 1 3.33 4 17.39
3+f aama £ dif. aan.
in aama f i dif. para. - 4 11.43 - - - - - -
Note, f ■ frequency of occurrencaa within a given group; % - percentage
of occurrencaa within a given group; 2, dif. aan. ■ combinations of 2
idea unita from different aentencea; 3+, dif* aan. - combinationa of 3
idea unita from different aentencea; 3+, aama £ dif. Ben. in dif. para.
* ■ combinationa of 3 or more idea unita from the aama sentence and from
different sentences in different paragraphs; 3+, aame £ dif. sen. in
same £ dif. para. - combinationa of 3 or more idea unita from the aame
sentence and from different sentences in the aame and in different
paragraphs.
As shown in Table 32, the results of the pre
intervention task show that the Experimental Group and
Control Group 1 produced more combinations of idea units
across paragraphs than Control Group 2 (Experimental Group
and Control Group 1 = 16.66%, Control Group 2 » 6.66%).
Table 32 also shows that all of the groups increased the
overall percentage of idea unit combinations across
paragraphs in the post-intervention summarization task.
However, Control Group l produced the highest percentage of
combinations across paragraphs (Control Group 1 « 50.00% vs.
Control Group 2 «* 34.78% vs. Experimental Group « 28*57%).
In addition, the results of the post-intervention
Total 2 16.66 10 28.57 3 16.66 5 50.00 2 6.66 8 34.78
141
summarization task show that Control Group 1 produced the
highest percentage of three or more idea unit combinations
from the same sentence and from different sentences in
different paragraphs (30.00%). The Experimental Group,
however, was the only group that produced three or more idea
unit combinations from the same sentence and from different
sentences within and across paragraphs (11.43%).
Idea unit reproduction strategies. Figure 9 displays
the frequencies of idea unit reproduction strategies, which
included copies, near copies, quotations, and paraphrases,
and Table 33 presents the frequencies and percentages of
copies, near copies, quotations, and paraphrases.
■ C o p ie s pro
□ C o p ie s post
■ N e a r copies pre
■ N e a r copies post
■ Q uotations pre
D E O Quotations post
■ P a ra p h ra se s pre
^ P a ra p h ra s e s post
experimental group control group 2
control group 1
Ficrure 9. Animal Biology: Frequencies of idea unit
reproduction strategies by group.
142
Table 33
Idea Unit: Reproduction Strategies in Animal Biology:
Frequency and Percentage bv Group
Reproduction
Strategies
Ettperiaental Oroup
Pre Post
t \ i \
Control
Pre
f %
Oroup 1
Post
f %
Control oroup 2
Pre Post
f % f %
Copies 2 16.66 6 17.14 5 27.77 4 40.00 9 30.00 3 13.04
Near copies 2 16.66 - ■ - 4 22.22 1 10.00 8 26.66 3 13.04
Quotations
- -
1 2.86 2 11.11 - - 2 6.66 1 4.34
Paraphrases 8 66.66 28 80.00 7 38.88 5 50.00 11 36.66 16 69.56
Total 12 99.98 35 100.00 18 99.98 10 100.00 30 99.98 23 99.98
Note, f - frequency of occurrences within a given group; % * percentage
of occurrences within a given group.
Table 33 shows that the Experimental Group paraphrased
more often than the other two groups in the pre-intervention
summarization task (66*66% vs. 38.33% Control Group 1,
36.66% Control Group 2). An increase in paraphrasing was
observed for the three groups, with Control Group 1
paraphrasing the least and copying the most. In addition,
all of the groups reduced the percentages of near copies
from the pre- to the post-intervention task (Experimental
Group - -16.66, Control Group 1 ■* -12.22, and Control Group
2 * -13.62). Nearly 17% of the idea units produced by the
Experimental Group in the pre-intervention task consisted of
copies. However, the Experimental Group copied less often
than Control Groups 1 (27.77%) and 2 (30.00%). The results
of the post-intervention task show that there was a very
small increase in the number of copies produced by the
Experimental Group (.48), a larger increase in the number of
143
copies produced by Control Group 1 (12.23) and a reduction
in the number of copies produced by Control Group 2
(-16.96).
Summary of Findings
The findings of this study indicate that the treatment
had a significant effect on the holistic performance of the
Experimental Group, which obtained significantly higher
holistic scores than those of Control Groups 1 and 2. The
treatment also had a significant effect on the performance
of the Experimental Group, as indicated by the number of
main ideas included in the summaries.
Following instruction, the Experimental summaries with
a sum-score of 9 and above were far more efficient than
Control Group 1 or Control Group 2 summaries. Moreover,
different patterns of strategy use were identified for the
three groups in the three content areas. For example, the
Experimental Group subjects differed from the subjects in
the two Control Groups in their gain in individual idea unit
production. By contrast, the percentage of individual idea
units produced by Control Group l subjects decreased in the
three content areas. Control Group 2 subjects exhibited a
very small increase in the production of individual idea
units— for Animal Biology .73— or did not produce any
144
individual idea units at all— for Health Science and
Cultural Anthropology.
The pattern of combinations across sentences within
paragraphs indicates that the Experimental subjects in
Health Science and Cultural Anthropology exhibited a gain
while the Experimental subjects in Animal Biology exhibited
a reduction. Control Group 1 subjects in Health Science and
Cultural Anthropology also gained in the number of
combinations across sentences within paragraphs. However,
their gain was much smaller than that exhibited by the
Experimental Group subjects. In contrast to the
Experimental Group and to Control Group l subjects, Control
Group 2 subjects in the three content areas exhibited a
reduction in the number of combinations across sentences
within paragraphs.
The patterns of combinations across paragraphs show
that Cultural Anthropology subjects in the Experimental
Group and in Control Groups 1 and 2 were equally inefficient
in combining information from the different sections in the
article. Moreover, Control Group 2 subjects in Health
Science exhibited the largest reduction in combinations
across paragraphs (-16.60), followed by a small reduction by
the Experimental Group (-2.90), and by a small gain by
Control Group 1 (2.70). All of the groups in Animal Biology
exhibited gains in combinations across paragraphs. However,
145
the results of this study reveal that Control Group 1
exhibited the highest gain (33.34), followed by Control
Group 2 (28.11), with the Experimental Group showing the
smallest gain (11.91).
Finally, the three groups also differed in their idea
unit reproduction strategies. The Experimental Group in
Health Science and Cultural Anthropology copied less often
in the post- than in the pre-intervention task. Although
the Experimental Group in Animal Biology exhibited a gain in
the number of copies produced, its gain was irrelevant
(.5%). In addition, the post-intervention results show that
the Experimental subjects in the three content areas
produced fewer copies and near copies— except for the
slightly higher percentage of near copies produced by the
Experimental subjects in Cultural Anthropology— than Control
Group 1 or 2 subjects. Chapter V provides a more detailed
summary of the findings, as well as a discussion of these as
they relate to prior research.
146
CHAPTER V
DISCUSSION OF FINDINGS
This chapter presents a summary of the findings
presented in Chapter IV, as well as an examination of the
features that differentiate the summary protocols produced
by the subjects in the Experimental Group and in Control
Groups 1 and 2. In examining the findings of this study,
attention will be paid to how the present results compare
with previous research.
The chapter is organized into four sections. The first
section provides an examination of the summary writers'
holistic performance. The second section presents an
analysis of the summary protocols in relation to the
students' production of main ideas and to the students'
summarization efficiency. The third section provides an
analysis of the strategies used to combine and to reproduce
the idea units contained in the main ideas. The fourth
section presents a discussion of explicit instruction as it
relates to the results presented in Chapter IV. The chapter
concludes with a summary of the discussion.
Holistic Performance
An Analysis of Covariance (ANCOVA) was conducted to
test for statistical differences in the holistic scores
attained by the Experimental Group and by Control Groups 1
147
and 2 on a summarization task. The statistical analysis
confirmed the group outcome differences for holistic
performance on the summarization activity. Of the two
covariates, the subjects' holistic scores on the pre
intervention task and the subjects' length of residence in
the United States, only the former was significant. The
fact that length of residence in the United States was not
significant may, at first glance, seem surprising. One
might think that length of residence in the U.S.— and,
therefore, attendance in U.S. schools— would have an impact
on the subjects' performance on an academic literacy task.
However, as indicated by the subjects, they participated in
a variety of U.S. school programs and there may have been a
great deal of variation in the quality of these programs.
Therefore, without more specific information, at this time,
it is impossible to pinpoint the reasons for this finding.
Post-hoc ANCOVAS were run to test for statistical
differences between the Experimental Group and Control Group
1, between Control Group 1 and Control Group 2, and between
the Experimental Group and Control Group 2. These tests
revealed that the Experimental Group's post-intervention
holistic scores were significantly higher than those
obtained by Control Groups 1 and 2.
The statistical analysis was expected to reveal
significant differences between the Experimental Group and
148
Control Group 1. Both groups were comparable on a number of
variables. For example, the Experimental Group and Control
Group 1 consisted of freshman subjects required to enroll In
pre-baccalaureate reading and writing courses. In addition,
an analysis of the subjects' records indicated that the
Experimental Group and Control Group 1 mainly consisted of
subjects in their second or third quarter of college
courses. Therefore, it could be assumed that these subjects
had not learned to shape their writing according to the
conventions of academic English discourse (Bartholomae,
1985).
By contrast, the statistically significant differences
in the post-intervention holistic scores attained by the
subjects in the Experimental Group (adjusted mean » 12.17)
and in Control Group 2 (adjusted mean “ 8.31) were not
expected. The subjects in Control Group 2 had already
completed the freshman English composition requirement and
were eligible to take the Writing Proficiency Examination
(WPE)— the CSLA's version of the CSU-mandated Graduation
Writing Assessment Requirement (GWAR). Additionally, the
subjects in Control Group 2 were more experienced than those
in Control Group 1 and in the Experimental Group because
they had been college students for a longer period of time.
In fact, 60% of the subjects in Control Group 2 were
sophomores and 20% were juniors. Because Control Group 2
149
subjects had been in college for a longer period of time
than Experimental or Control Group 1 subjects, they had more
experience with the types of academic literacy tasks
required to fulfill assignments in different disciplines.
The findings of this investigation also indicate that
there was a change in the performance of the Experimental
Group, as indicated by the number of subjects in the group
who attained a sum-score of 9 and above in the pre- and
post-intervention tasks. Specifically, the results of the
pre- and post-intervention tasks show that nine of the 20
subjects in the Experimental Group moved from the 8-or-below
to the 9-and-above holistic score category, and only two
subjects in Control Group 1 and one subject in Control Group
2 moved from the 8-or-below to the 9-and-above category. As
expected, the performance of the subjects in Control Groups
1 and 2 remained constant across tasks. Good and poor
summarizers in Control Groups 1 and 2 did not show major
improvements or decreases in their scores, as indicated by
the number of subjects who moved from 8-or-below to the 9-
and-above holistic score category. The intervention was
effective in helping nearly half of the subjects in the
Experimental Group obtain higher holistic scores, as shown
by the number of subjects who moved from the 8-or-below to
the 9-and-above category.
150
As already indicated, the adjusted mean scores were
12.17 for the Experimental Group, 8.13 for Control Group l,
and 8.31 for Control Group 2. A mean of 12.17 was
indicative of summarizers who demonstrated minimal
competence in summary writing at the content, rhetorical,
and syntactic levels. In contrast, the lower means attained
by both Control Groups revealed that the summarizers
demonstrated some developing competence in summary writing,
but the summaries remained flawed at either the content,
rhetorical, or syntactic level.
An evaluation of features of the summary protocols
suggests three qualitative differences in the summaries
produced by the three groups. First, in the pre- and post
intervention tasks, Control Group 2 subjects maintained an
objective tone while Control Group 1 subjects responded to
the source text in a personal way. As was the case for
Control Group 1 subjects, the subjects in the Experimental
Group failed to maintain an objective tone in the pre
intervention task; by contrast, they were successful in
displaying objectivity in the post-intervention task. This
objectivity in tone may have contributed to the Experimental
subjects' increased holistic scores on the post-intervention
summarization task, supporting earlier findings by Connor
and NcCagg (1987).
151
Lack of objectivity, as indicated by the writers'
personal response to the text, was observed mainly among
Health Science summary writers, with some instances in the
Cultural Anthropology and Animal Biology subjects. The
following is an example showing how a male writer reacted to
the source text by expressing his opinion about the topic:
Why can't physicians just look at everyone else as
human beings instead of discrimaniting, heart
disease doesn't. (#05)
As will be shown in the examples that follow, other writers
chose to respond explicitly to the source text:
Why I feel surprised is because according with the
article heart disease is a man's problem but about
what I read. (#19)
What I basically understood from the article...(#17)
These results are consistent with Basham's (1986) findings,
who argued that unacculturated Alaska Native (AN) summary
writers also responded explicitly to a source text. More
importantly, these results support previous findings by
Basham, Ray, and Whalley (1993) regarding the "orientation
toward topic" (p. 310) exhibited by a group of bilingual
Latino students in university contexts. Still, the
"orientation toward the topic" in this study was only
displayed by unacculturated readers and writers, and not by
the "able" subjects in Control Group 2.
The lack of objectivity displayed by the subjects in
the Experimental Group, as well as by the subjects in
152
Control Group 1 in the pre-intervention summarization task
may have been the result of one or a combination of two
factors. The first factor involves the nature of the Health
Science article. Because the Health Science article dealt
with an issue of more immediate relevance, the subjects may
have been prompted to respond in a personal way. By
contrast, the "foreigness" (Johns, 1991c) of the topics of
murder in different societies and winter depression,
presented in the Cultural Anthropology and Animal Biology
articles respectively, may have prevented the subjects from
tapping into their background knowledge (Carrell &
Eisterhold, 1988) and may have caused them to become less
personally involved with the reading (Coady, 1979).
The second factor that may have affected objectivity in
the pre-intervention summary protocols is the writers' lack
of acculturation into the demands of university literacy
tasks. This explanation would be supported by the detached
tone displayed by Control Group 2, consisting of sophomore
and junior subjects. In any case, neither of the two
explanations— the foreigness of the topic or the subjects'
lack of acculturation to academic demands— support
conclusions reached by Basham (1986). Specifically, in her
study of summary protocols produced by AN college subjects,
Basham (1986) attributed her subjects' lack of objectivity
153
to "... cultural attitudes about the nature of texts and
interpersonal relations1 ' (p. 174).
The second feature that differentiates the three
groups' protocols involves the strategies that the subjects
employed to open their summaries. In general, the results
of the pre- and post-intervention task show that summary
writers in Control Group 2 began with an explication of the
pragmatic condition of the task (Connor 6 McCagg, 1987).
Two examples from the corpus illustrate this point:
This article entitled "Here Comes the Sun" describes a
disorder called "Seasonal Affective Disorder (SAD).
(#54)
The article I read, "The Safer Sex? Probing a
cardiac gender gap," by Kathy A. Fackelmann was about
the risks of women getting a heart attack and many
researchers who did studies to try to find findings
that may be different conditions that are found in men
with heart problems. (#62)
Most of the pre-intervention summary writers in the
Experimental Group and in Control Group 1 failed to begin
with an explication of the pragmatic condition of the task.
For example,
Physicians in the 90's have largely ignored women
who complain of chest pains which is a big potential
sign of heart attack. (#06)
Bruce Bower describes a women who is lieing asleep,
then a light bult being place near her bed. (#42)
In general, writers who did not open their summaries with an
explication of the pragmatic condition of the task either
failed to attribute the ideas presented to an author other
154
than the summary writer (protocol #06) or made assumptions
about the reader's shared knowledge of the topic (protocol
#42).
An interesting change was found in the Experimental
Group's post-intervention summary protocols. In general,
the first sentence in the summaries included an explication
of the pragmatic condition of the task (see italicized
example).
In The Safer Sex?" the author talks about who is
more likely to die from a myocardial infarction,
whether males or females. (#05)
In the article "Mere Comes the Sun" the author
reports research done on winter depression and how
it affects people. (#37)
Although, in the post-intervention task, some of the
subjects in Control Group 1 opened their summaries by
stating the pragmatic condition of the task, most writers in
this group failed to attribute the ideas to the author of
the article.
The third qualitative difference identified in the
summaries produced by the three groups involves the
inclusion of macropropositions. The results of this
investigation show that, in the post-intervention
summarization task, the Experimental subjects attempted to
begin their summaries with macropropositions which would
give the gist of the article. This finding is particularly
interesting in light of Ll and L2 research that suggests
1SS
that summarization-strategy training does not commonly
result in the ability to create macropropositions
(Bensoussan & Kreindler, 1986; Brown & Day, 1983; Day, 1980;
Garner & HcCaleb, 1985; Hare & Borchardt, 1984; Johns &
Hayes, 1990). Although this investigation cannot definitely
ascertain that summarization instruction resulted in the
Experimental subjects' increased ability to create
macropropositions, it can be argued that the summaries
written by the Experimental subjects exhibited a trend
toward including macropositions more often than did the
summaries written by Control Group 1 subjects.
This trend was observed in the summaries of the
Experimental subjects in Health Science and Animal Biology
only. It may well have been the case that the content
instructors played an important role in this respect. In
the course of their lectures, the Health Science and Animal
Biology instructors emphasized the importance of opening any
reading-to-write assignment with a generalization and
focusing on the author's message. Nevertheless, the trend
toward macropositions cannot be attributed solely to the
content instructors; otherwise, Control Group 1 subjects,
who attended the same lectures, would also have produced
macropropositions. For the Health Science and the Animal
Biology subjects, the combination of the content
156
instructors' emphasis and the summarization-strategy
training must have influenced their performance.
Main Ideas and Summarization Efficiency
Prior research has shown that "able" and "less able"
readers differ in their sensitivity to the importance of
textual ideas (Spivey, 1983; Winograd, 1984). The results
of the pre-intervention summarization task are consistent
with these findings. Specifically, Control Group 2,
consisting of "able" readers, recognized a higher number of
main ideas than the Experimental Group and Control Group 1.
Summarization-strategy training was effective to the extent
that, following instruction, the Experimental Group
reproduced the highest number of main ideas among the three
groups (adjusted mean, Experimental Group - 3.85; adjusted
mean, Control Group 1 - 1.97; adjusted mean, Control Group 2
- 2.37). In fact, an ANCOVA revealed that there were
significant differences among the Experimental Group and
Control Groups 1 and 2 in their performance on the
summarization task, as indicated by the number of main ideas
present in the subjects' summaries. Moreover, a post-hoc
ANCOVA confirmed statistical differences between the
Experimental Group and Control Group 1. Although the
Experimental Group produced a higher number of main ideas
157
than Control Group 2, the difference did not reach
statistical significance.
While these results indicate that there was a
significant improvement in the Experimental Group's
sensitivity to the importance of textual ideas, the mean
number of main ideas recognized not only by the Experimental
Group but also by Control Groups 1 and 2 was low (adjusted
mean, Experimental Group * 3.85, adjusted mean, Control
Group 1 « 1.97, adjusted mean Control Group 2 ■ 2.37).
This deficiency in performance was probably not the
result of a lack of awareness of the requirements of the
summarization task; as shown in Chapter III, the subjects in
this study were aware that summary writing involves the
identification of main ideas and the subsequent reduction of
information (Garner, 1985; Rinaudo, 1993; Winograd, 1984).
The deficiency in performance found in this study is
consistent with the results of Rinaudo (1993), who found
that, although first-year university subjects understood the
requirements of a summarization task, 50% of her subjects
failed to identify 50% of the main ideas in the source text.
The results of this study, however, contradict earlier
findings by Garner (1985), whose college subjects identified
nearly all of the main ideas in the source text. These
contradictory findings seem to be the result of the
characteristics of the subjects included in the studies, and
158
to a lesser extent, the nature of the source text. First,
Garner's subjects, college students enrolled in an
undergraduate reading methods class, may have been more
"strategic" readers than subjects registered in general
education courses, second, the relative effects of topic
difficulty and text structure in Garner's (1985) study are
unknown. As indicated by Garner herself, the number of main
ideas contained in the article read by her subjects were
four, as opposed to the 19, 17, and 14 main ideas contained
in the articles used in this investigation. Therefore, it
is not inappropriate to assume that Garner's subjects may
have had an easier time identifying the main ideas.
The deficiency in performance, as indicated by the
number of main ideas included in the subjects' summaries,
may have resulted from the subjects' inability to identify
important information or from the effect of the structure of
the source texts on the subjects' ability to recall
information. The students may have been prevented from
identifying a greater number of main ideas due to the
partially descriptive nature of the source texts, which have
been found to be the least effective in facilitating
information recall (Carrell, 1992; Meyer, Brandth, & Bluth,
1980; Meyer & Freedle, 1984).
Although the treatment effect, as indicated by the
number of main ideas included in the summaries, was found to
159
be statistically significant (p = .003), the subjects might
have benefited even more greatly from strategy instruction
that combined both summarization and text structure
training. Text structure training could have provided the
subjects with an understanding of how to identify the main
ideas in the source texts and, in turn, this training could
have resulted in the identification of a greater number of
main ideas. This contention is supported by LI and L2
reading research, which has shown that text structure
training enhances reading comprehension and recall of
information for expository text (Armbuster, 1991; Armbuster
& Anderson, 1980; Armbuster, Anderson, & Ostertag, 1987;
Berkowitz, 1986; Carrell, 1985, 1992; Gordon, 1990; Grabe,
1993; Johns, 1985; Sjostrom & Hare; 1984).
Besides producing a greater number of main ideas, the
Experimental subjects also gained in the number of adjacent
main ideas produced. In addition, the results of the post
intervention summarization task show that the Experimental
subjects in Cultural Anthropology and Animal Biology
connected adjacent main ideas via conjunctive relations more
often than they did in the pre-intervention task. In
connecting two main ideas in the post-intervention task, the
Experimental subjects in the three content areas relied
mainly on additive, adversative, and causal conjunctions,
though to differing degrees. For example, Animal Biology
160
subjects relied on adversative and additive conjunctions
equally and, Cultural Anthropology subjects used adversative
conjunctions more often than additive or temporal
conjunctions. Finally, Health Science subjects used
additive conjunctions more often than causal or adversative
conjunctions.
control Group x subjects in Health Science and Animal
Biology relied on additive conjunctions heavily. Control
Group 1 and 2 subjects in Cultural Anthropology did not
produce any adjacent main ideas; on the other hand, Control
Group 2 subjects in Health Science connected two main ideas
with additive conjunctions; similarly, Animal Biology
subjects relied on additive and causal conjunctions in the
same degree.
In sum, all of the groups seemed to rely heavily on
additive and adversative conjunctions in connecting two
adjacent main ideas. Although this investigation did not
measure the writers' level of literacy in Spanish, the
findings of this study are partially supported in research
by Montafto-Harmon (1988), who concluded that Mexican Spanish
writers organize written text via additive relationships.
Therefore, Montafto-Harmon's findings may help explain the
frequent usage of additive conjunctions among the three
groups that participated in this study.
161
Prior to summarization instruction, the summaries
produced by Experimental subjects— except for the summaries
produced by Health Science subjects— were similar to the
summaries produced by Control Group 1 subjects in that the
central information contained in the summaries was very
limited. By contrast, following instruction, the results of
the ANCOVA show that the summaries written by the
Experimental subjects in the three content areas were
comparable to those produced by Control Group 2 subjects.
The summary writers focused more on the central ideas
contained in the articles, and excluded irrelevant
information. This increased focus on important information
is an indication of the higher quality of the content
present in the summaries produced by the subjects in the
Experimental Group and in Control Group 2. In turn, content
quality, found to be of primary concern for content
instructors (Bridgeman & Carlson, 1984; Rusikoff, 1994), may
have prompted the holistic raters to assign higher holistic
scores to Experimental subjects. Content quality, as
indicated by the number of main ideas included in the
summaries, however, may have contributed to the higher
holistic scores attained by the Experimental subjects; but
content quality could not have been the only factor that
accounted for higher holistic scores. If this had been the
case, the ANCOVA would not have demonstrated significant
162
differences in the holistic scores attained by the subjects
in the Experimental Group and in Control Group 2.
The differences in the holistic scores attained by the
three groups may be, at least partially, accounted for by
the differences in summarization efficiency, as indicated by
the mean ratio of the number of main ideas to the number of
words included in the summary protocols. As expected,
Control Group 2's pre-intervention summaries having a sum-
score of 9 and above were the most efficient. This means
that Control Group 2 summaries contained fewer words than
the other two groups' summaries, but expressed a greater
number of main ideas. In addition, pre-intervention results
show that the summaries produced by the Experimental Group
having attained a sum-score of 9 and above did not exhibit
major differences in summarization efficiency from those
shown by the summaries produced by Control Group 2 having
attained a sum-score of 8 and below. This lower
summarization efficiency for summary protocols with higher
holistic scores may have been the result of the subjects'
inability to be succinct (Garner, 1985). Furthermore, the
mean efficiency ratio may have been skewed by the wide range
of main ideas and words included in the summary protocols.
The lower summarization efficiency ratio attained by
summary writers in the Experimental Group and in Control
Group 1 in the pre-intervention summarization task did not
163
result from the subjects' lack of awareness that summarizing
Involves reproducing important information at the expense of
unimportant details. As noted elsewhere in this chapter,
prior to summarization instruction, the subjects were well
aware of the requirements of a summarization task.
Post-intervention results show practically no change in
efficiency for those summaries written by Control Group 2
having a sum-score of 9 and above. On the other hand, the
summaries produced by the Experimental Group receiving a
sum-score of 9 and above were far more efficient than those
produced by control Group 2 in the same category. In fact,
the summarization efficiency for the Experimental Group
nearly tripled from that in the pre-intervention task and
was nearly twice as large as the summarization efficiency
for Control Group 2 in the post-intervention task. Although
instruction was expected to result in improved summarization
efficiency for the Experimental Group, such a dramatic
increase was not expected. These results confirm that the
Experimental Group not only produced summaries that
contained a greater number of important ideas but also that
they excluded information that was irrelevant. The
Experimental Group may have, consciously or not, engaged in
a process of tree trimming, to use Rumelhart's (1977b) term,
meaning that content which was at the lower level of the
tree structure was discarded.
164
A second ratio of summarization efficiency, as
indicated by the ratio of the number of main ideas to the
number of idea units used to express the main ideas, was
calculated for each of the three content areas. Prior to
this investigation, it was hypothesized that good summary
writers would use fewer idea units to express the main ideas
in the articles. However, the pre- and post-intervention
summaries produced by the subjects in the Experimental Group
and in Control Groups l and 2 in the three content areas
displayed a very wide range of mean efficiency ratios, and
no clear patterns emerged. A probable reason for this lack
of clear patterns may have been the choice of the mean,
which may have masked real differences among the groups. It
remains to be seen whether the main idea efficiency
calculation renders more useful results with highly
proficient readers and writers who may, purposefully, intend
to use fewer words and, probably by extension, fewer idea
units to express main ideas.
Writing Strategies
As stated in Chapter IV, the summary writers in this
study employed a number of strategies to reproduce the idea
units contained in the main ideas. This section will focus
on two main areas, including the strategies used to combine
16S
information within and across paragraphs and the strategies
used to reproduce written text.
Combination Strategies
The three groups differed in the strategies used to
combine the main ideas reproduced in the summaries.
Specifically, the Experimental Group differed from the two
Control Groups in its gain in the percentage of individual
idea unit produced in the three content areas. By contrast,
the percentage of individual idea units produced by Control
Group 1 decreased in the three content areas. Control Group
2 either exhibited a minimal gain in the production of
individual idea units (.73 for Animal Biology) or did not
produce any individual idea units at all (for Health Science
and Cultural Anthropology).
The gain in individual idea units exhibited by the
Experimental Group subjects may be accounted for by the
demands of the summarization task. Once the subjects in the
Experimental Group successfully identified a main idea
consisting of only one idea unit, they may have felt
constrained by time limitations, or their limited ability to
combine two or more idea units may have been the result of
the students' inability to engage in abstraction operations,
limiting them to the use of local processing strategies
(Eskey & Grabe, 1988). When producing individual idea
166
units, Experimental subjects, consciously or not, directed
their attention at the word and sentence levels rather than
at the discourse level.
The reduction in the production of individual idea
units for the subjects in Control Group 1 cannot be
accounted for by a gain in idea unit combinations within
paragraphs. The subjects in control Group l experienced a
sharp drop in idea unit combinations across sentences within
paragraphs (mean gain - -3.22), mainly accounted for by a
decrease in the production of combinations across sentences
within paragraphs exhibited by the Animal Biology group.
Neither can a reduction in the production of individual idea
units exhibited by Control Group l be accounted for by a
gain in combinations across paragraphs since only the Animal
Biology subjects improved its performance.
These results indicate that Control Group 1 subjects in
Cultural Anthropology and Health Science employed local
summarization strategies more often than the global
strategies characteristic of good readers (Eskey & Grabe,
1988). In summarizing, the subjects may have focused on
matters related to time constraint and grammatical accuracy,
increasing the cognitive load on summary writers, and
probably limiting their ability to combine ideas presented
in different sections of the articles.
167
In contrast to Control Group 1 subjects in Cultural
Anthropology and Health Science, Control Group 1 subjects in
Animal Biology seemed to perform as proficient readers in
the post-intervention task. By combining information from
different paragraphs, Control Group 1 subjects in Animal
Biology showed that they were able to read the text globally
and engage in abstraction operations.
Control Group 2 subjects also experienced a sharp drop
in idea unit combinations across sentences within paragraphs
(mean reduction “ 12) . A high number of idea units
produced by Control Group 2 subjects in the pre-intervention
task (23.33% in Animal Biology, 36.66% in Cultural
Anthropology, and 50% in Health Science) consisted of
combinations across sentences within paragraphs. As had
been expected, pre-intervention results showed that Control
Group 2 subjects, characterized as "able" readers, were able
to go beyond the sentence level and to focus on the
paragraph level.
The results for combinations across paragraphs for
Control Group 2 subjects were different. For example,
Cultural Anthropology subjects did not produce any
combinations across paragraphs either in the pre- or in the
post-intervention task. Therefore, the mean gain in
combinations across paragraphs for the three groups was
affected by the poor performance of Cultural Anthropology
168
subjects (mean gain » 3.83). Additionally, this mean is
highly skewed due to the reduction in idea unit combinations
across paragraphs for Health Science (from 16.6% in the pre
intervention task to 0% in the post-intervention task).
Therefore, the gain in combinations across paragraphs for
Control Group 2 subjects can be accounted for mainly by the
subjects in Animal Biology.
The reduction in individual idea unit production for
Control Group 2 subjects is difficult to understand because
it was accompanied by a decrease in combinations across
sentences within paragraphs and by a very low increase in
idea unit combinations across paragraphs. Control Group 2
subjects appear to have been affected by a reduction in the
production of idea units in general (from 47 in the pre- to
40 in the post-intervention task), followed by a
rearrangement in how the idea units were combined.
Control Group 2 subjects were expected to combine more
information across paragraphs. However, contrary to
expectations, the results of the post-intervention task
suggest that Control Group 2 subjects did not seem to be
able to go beyond the paragraph level and abstract global
meaning. Although the ability to combine information across
paragraphs is a late-developing skill (JohnB, 1985, Johns &
Hayes, 1990), the fact that Control. Group 2 subjects
169
experienced difficulty in combining information across
paragraphs was surprising.
It is unclear whether the experimental condition of the
task may have affected the subjects when completing the
post-intervention task; nevertheless, at the time of the
administration of the post-intervention task, this did not
seem to be the case. The subjects seemed to have been
highly motivated by the fact that, for completing the task,
they obtained points that counted toward the final grade in
each of the three general education courses. What may have
happened is that the small n» particularly in Cultural
Anthropology and Health Science, may have skewed the results
of this study. In addition, the "foreigness" of the article
employed in the Cultural Anthropology class may be a likely
explanation for the lack of combinations exhibited by
Experimental, Control Group 1, and Control Group 2 subjects
in Cultural Anthropology.
At the same time that the Experimental subjects
experienced an increase in individual idea unit production,
there was a gain in idea unit combinations across sentences
within paragraphs for Experimental subjects in Health
Science (13.26) and Cultural Anthropology (51.29), and a
reduction in combinations for Experimental subjects in
Animal Biology (8.09).
170
The results of the pre-intervention task show that the
Experimental Group subjects produced the lowest percentage
of idea unit combinations across sentences within paragraphs
(10.5%), with Control Group 2 subjects producing the highest
(36.66%). Post-intervention results show that the
Experimental subjects experienced the largest mean gain
(18.82 vs. Control Group 2 « 12 vs. Control Group l «*
-3.2).
Nearly 10% of the idea units produced by the
Experimental Group in the pre-intervention summarization
task consisted of combinations across paragraphs. This
percentage was higher than those for Control Group 2 (7.75%)
and Control Group 1 (5.55%). Post-intervention results show
that, although the Experimental Group gained in combinations
across paragraphs, mainly accounted for by the Animal
Biology group, Control Group 1 experienced the largest gain,
mainly accounted for by the Animal Biology group. As
already indicated, the Experimental Group, whose post
intervention summaries were the most efficient, as indicated
by the mean ratio of the number of main ideas to the number
of words included in the summaries, may have had a difficult
time producing more ideas in fewer words, combining ideas
from different sections of the paragraphs, and avoiding
plagiarism.
171
Reproduction Strategies
The Experimental subjects in Health science and
Cultural Anthropology produced fewer copies in the post-
than in the pre-intervention summarization-task. In
contrast, the Experimental subjects in Animal Biology copied
more, but by less than .5%. Additionally, the Experimental
subjects in the three content areas produced fewer near
copies and more paraphrases in the post- than in the pre
intervention task.
The Experimental subjects in the three content areas,
however, exhibited different paraphrasing patterns. The
Experimental subjects in Health Science exhibited the
highest gain in paraphrasing (42.66), immediately followed
by the subjects in Cultural Anthropology (41.66), with the
Experimental subjects in Animal Biology (13.34) exhibiting
the lowest gain. The low gain in paraphrasing by the
subjects in Animal Biology could be explained by the fact
that nearly 67% of the idea units contained in the main
ideas in the pre-intervention task were paraphrased. At
first glance, this finding was rather surprising, given that
the three articles had similar text structures and that the
topic presented in the Animal Biology article was thought to
be more "foreign" than that in the Health Science article.
Therefore, at this time, the only available explanation for
the higher number of paraphrases produced by Experimental
172
subjects in Animal Biology in the pre-intervention task is
that these subjects were more efficient paraphrasers from
the outset.
It is important to note that the post-intervention task
results show that, as Experimental subjects produced fewer
copies (except for the .5 increase for Animal Biology) and
near copies, the percentages for paraphrasing increased.
These findings are consistent with those of Campbell (1990),
who found that, as the use of inappropriate reproduction
strategies such as copying and near copying was reduced,
subjects resorted to appropriate reproduction strategies,
such as paraphrasing.
Control Group 2 subjects in Health Science and Animal
Biology produced fewer copies and near copies and more
paraphrases in the post- than in the pre-intervention task.
Post-intervention results show that over 69% of the idea
units contained in the main ideas produced by Control Group
2 subjects in Health Science and Animal Biology were
paraphrased. Still, Control Group 2 subjects in the three
content areas paraphrased less often than the Experimental
Group subjects. In addition, the mean gain in paraphrases
by the subjects in Control Group 2 in the three content
areas was 1.29, as opposed to the 32.55 exhibited by the
subjects in the Experimental Group.
173
The post-intervention results for the subjects in
Control Group 2 in Health Science and Animal Biology
indicate that the "able" group functioned as such; Control
Group 2 subjects employed strategies typical of proficient
readers and writers. They found synonyms for content words
and phrases, a skill that appears late and requires the
reader not only to understand written text but also to have
a large vocabulary (Johns & Mayes, 1990).
More importantly, the findings of this study show that
the post-intervention reproduction strategies exhibited by
the subjects in the Experimental Group in the three content
areas did not differ from those employed by the subjects in
Control Group 2. Clearly, instruction resulted in the
higher use of appropriate reproduction strategies,
characteristic of better LI and L2 readers (Brown 6 Day,
1983; Hare & Borchardt, 1984; Johns, 1985; Johns & Mayes,
1990; Winograd, 1984).
In contrast to the subjects in the Experimental Group
in the three content areas and to the subjects in Control
Group 2 in Animal Biology and in Health Science, the
subjects in Control Group 1 copied more often in the post-
than in the pre-intervention summarization task. Control
Group 1 subjects also copied more and paraphrased less than
did the subjects in the Experimental Group. However, like
the subjects in the Experimental Group and in Control Group
174
2, Control Group 1 subjects exhibited a gain in paraphrasing
(mean gain « = 12.12). Still, Control Group 1 subjects seemed
to rely on the copy-delete strategy more heavily than did
the Experimental subjects. These findings support earlier
studies in Ll and L2 reading research (e.g., Brown & Day,
1983; Garner, 1985; Johns, 1985; Johns & Mayes, 1990;
Winograd, 1984), which have identified developmental
differences in the acquisition of summarization strategies.
The mean gain in paraphrasing exhibited by the three
groups (Experimental Group - 32.5 vs. Control Group l “
12.12 vs. Control Group 2 * 1.29) could have resulted from
the gain in background knowledge in the specific academic
content area. Lack of background knowledge during the pre
intervention task may have affected the subjects' ability to
manipulate content-specific text (Alderson & Urquhart, 1984;
Carrell & Eisterhold, 1988). At the time of the poBt-
intervention task, the content knowledge gained throughout
the term may have lowered the readers' cognitive load,
leading them to understand the content better and to rely
less on copying.
However, as previously shown, while the three groups
produced more paraphrases in the post-intervention task, the
Experimental Group made the greatest gains. A number of
explanations can be given for this finding. For example, as
a result of instruction, the Experimental Group may have
175
improved in its command of the language, in its ability to
paraphrase written text, or in its awareness of the
inappropriateness of plagiarism.
As already stated, the subjects in the Experimental
Group and in Control Groups 1 and 2 (in Animal Biology and
Health Science) produced fewer instances of near copies in
the post- than in the pre-intervention task. In addition,
for the subjects in the Experimental Group, there appeared
to be a trend toward more near copies and fewer paraphrases
in the pre-intervention task, and a trend toward more
paraphrases and fewer near copies in the post-intervention
task. With the information available, it cannot be assumed
that the subjects in this study consciously intended to
avoid copying. Neither is it clear whether, given more time
to complete the in-class summarization task, the subjects
would have produced more paraphrases and fewer near copies.
Few instances of quotations were observed for the three
groups in the three content areas across tasks. However,
good and poor summarizers, as measured by the protocols'
holistic scores, employed different strategies to integrate
quotations in the summary protocols. For example, summary
protocols with high holistic scores included quotations
introduced by discourse markers such as according to.
Following is an example from the corpus:
176
According to Fackelroann, the idea that "heart
disease is a man's problem" is a stereotype
because ... (#07)
t
The summary protocols of unsuccessful summarizers, measured
by the subjects' low holistic scores, were characterized by
the inclusion of quotations that were not integrated in the
summaries. The following example from the corpus
illustrates this point.
"The Safer Sex?" by Kathy A. Fackelmann is an
article on the different affects heart disease has
on men and women. "Host people know that heart
disease reppresents the number-one killer of men in the
U.S. but many don't realize that it's also the number
one killer of women." (#05)
While successful summarizers quoted short segments from the
Bource text, unsuccessful summarizers tended to quote long
chunks of information from the source text. These subjects
may have been aware of the inappropriateness of plagiarism
and, once they identified important information, they
proceeded to quote it instead of copying it directly. In
addition, some subjects may have thought that their writing
style would never match that of a published author, and this
may have led the writers to quote instead of paraphrase.
Explicit Instruction
As indicated in Chapter III, the Experimental subjects
learned how to summarize through explicit instruction.
Explicit instruction included models, scaffolds, and guided
and supported practice. In addition, explicit instruction
177
required subjects to give and receive feedback for their
summaries, and withdrew teacher and student support as the
subjects became more competent summary writers.
The results of this investigation revealed that, with a
short intervention, subjects in the Experimental Group
obtained significantly higher post-intervention holistic
scores than the subjects in the two Control Groups. In
addition, the Experimental subjects produced a significantly
higher number of main ideas than subjects in Control Group
1.
The "underprepared" subjects involved in this study
benefited from explicit instruction in summarization
strategies. As supported by L2 acquisition research
(Hulstijn & Schmidt, 1994, for example), explicit
instruction facilitated the acquisition of a number of
summarization strategies over others. Specifically, the
Experimental subjects identified main ideas and deleted
redundant and minor details, and the summary protocols
written by the Experimental subjects included more
macropropositions than the summaries produced by Control
subjects. The Experimental subjects, however, had a
difficult time combining information from different sections
of the articles. In addition, although the summaries
produced by Experimental subjects received significantly
higher holistic scores than those attained by the summaries
178
produced by the two Control Groups, the summary protocols
written by Experimental subjects included many spelling and
grammatical mistakes. It is not clear whether, given more
time to revise their summaries, these subjects would have
produced better quality summaries; given the characteristics
of some of the subjects' mistakes, it is doubtful that this
would have been the case.
Although this study did not formally measure the
subjects' attitudes toward explicit instruction, it should
be observed that the Experimental subjects reported positive
reactions to the summarization-strategy training program.
As indicated by many of the subjects in this study, explicit
instruction made them aware of the elements of academic
summaries. In this respect, explicit instruction may have
played the role of a "focusing device for student attention"
(Hulstijn & Schmidt, 1994, p. 9).
Summarv_of Discussion, of Findings
As noted in this chapter, the statistical analyses in
this study revealed that the treatment was effective. The
results of an ANCOVA indicated that the holistic performance
of the "underprepared" L2 readers in the treatment group,
following treatment, was significantly superior to that of
both the "underprepared" and the "able" L2 readers without
treatment. A second ANCOVA revealed that the writing of the
179
treatment group included a significantly higher number of
main ideas than the "underprepared" subjects in Control
Group 1. Moreover, the post-intervention summaries produced
by the Experimental subjects were more efficient than those
produced by subjects in Control Groups 1 and 2.
In addition, there were qualitative differences in the
summary protocols produced by the subjects in the
Experimental Group and in Control Group 1. similar to
control Group 2, consisting of "able" readers, the
Experimental Group subjects produced post-intervention
summaries that had an objective tone, began with an
explication of the pragmatic condition of the task, and
presented more macropropositions.
The results of this study also revealed that the
Experimental subjects modified, to some degree, the
strategies used to combine and reproduce the main ideas in
the source text. Following summarization instruction,
rather than reading at the discourse level, the Experimental
subjects appeared to utilize local strategies when they
reproduced individual idea units. Nevertheless, the fact
that the Experimental subjects gained in combinations across
sentences within paragraphs is an indication that the
Experimental subjects were moving toward the use of more
global strategies. Zn this respect, however, summarization-
strategy training was not entirely successful in helping
iso
Experimental subjects to use abstraction operations
consistently and to combine a higher percentage of idea
units across paragraphs. In contrast, following
instruction, the Experimental subjects were successful in
reducing the number of copies, near copies, and quotations
and in increasing the use of paraphrases. In this area, the
Experimental subjects acted much like the "able" readers in
Control Group 2. Experimental subjects avoided using the
"copy-delete" strategy, characteristic of "underprepared"
summarizers, and identified in the Control Group 1 subjects.
Summarization instruction, in this study, included
content-specific materials and explicit training. Given the
impressive results obtained after a short intervention of
eight hours, one could speculate what the treatment effect
would have been if the intervention had been longer. A
longer intervention might have resulted in the
identification of a higher number of main ideas and in the
more frequent use of global reading strategies.
As explained in this chapter, the "underprepared"
subjects were not expected to attain significantly higher
holistic scores than the "able" subjects. Nonetheless, it
could be questioned whether the "able" subjects were truly
"able." Based on the holistic quality of the summaries
produced, the "able".subjects did not perform as would have
been expected. However, the fact that they had fulfilled
18!
the Freshman composition requirement entitled them to take
the WPE— the CSLAls version of the CSU-mandated GWAR— even
though they still appeared to be, to some extent, unable to
use strategies that characterize "able" readers
consistently. The "able" subjects, much like the
"underprepared" subjects in this investigation, would have
probably benefited from cognitive strategy instruction
anchored in the demands of the real curriculum.
182
CHAPTER VI
SUMMARY OF THE STUDY, CONCLUSIONS, AND RECOMMENDATIONS
This chapter is organized into three main sections.
The first section presents a summary of the study, including
its purpose and a summary of the findings. The second
section provides conclusions based on the findings of this
investigation. The last section presents recommendations
for classroom pedagogy and for further research.
Summary of the Study
Purpose of the Study
As indicated in Chapter I, the purpose of this study
was two-fold. From a narrow perspective, the purpose of
this study was to determine whether "underprepared" non
native speakers of English (NNS), university-level native
Spanish-speaking (NSS) freshmen would benefit from explicit,
content-based instruction in summary writing in an adjunct
class. From a broader perspective, the purpose of this
study was to determine the extent to which instructional
strategies that had been successful with first language (LI)
students would be applicable to "underprepared" second
language (L2) freshmen.
183
Summary of Findings
The findings of this study revealed that explicit
instruction in summary writing in the content areas proved
to be successful with "underprepared" NSS subjects. Prior
to instruction, the summary writers in the Experimental
Group and in Control Group l shared a number of
characteristics which are typical of inexperienced or
"underprepared" LI and L2 readers and writers.
Specifically, the summary writers attained low holistic
scores, did not maintain an objective tone in the summaries,
and did not open their summaries with statements of the
pragmatic condition of the task. In addition, the writers
employed inappropriate reproduction strategies such as
copying and nearly copying more often than appropriate
reproduction strategies such as paraphrasing.
Following summarization-strategy instruction, a number
of changes, as indicated by the results of the statistical
and qualitative analyses, were observed in the summaries
produced by the subjects in the Experimental Group. The
Experimental subjects behaved much more like the "able" L2
readers in Control Group 2 than like the "underprepared" L2
readers in Control Group 1. This section will provide a
summary of the main findings of this study.
First, as indicated in Chapter IV, the results of an
ANCOVA statistically confirmed the group outcome differences
184
for the summarization task. Post-hoc ANCOVA tests revealed
that the summaries written by the Experimental Group
attained significantly higher holistic scores than the
summaries written by Control Group 1, consisting of
"underprepared” L2 readers, and by Control Group 2,
consisting of "able" L2 readers.
Second, the results of another ANCOVA showed that the
treatment effect was significant for the number of main
ideas included in the summaries. A post-hoc ANCOVA revealed
that the number of main ideas reproduced by the Experimental
Group was significantly greater than the number of main
ideas reproduced by Control Group 1. Although the
Experimental Group reproduced a greater number of main ideas
than Control Group 2, the differences between both groups
did not reach statistical significance.
Third, a large difference was identified between the
mean number of main ideas reproduced by the Experimental
Group following instruction (adjusted mean » 3.85) and the
mean number of main ideas present in the articles (mean -
16.66). Although the Experimental Group became more
sensitive to the importance of textual ideas by over 100%,
its performance, as measured by the number of main ideas
reproduced, was marginal at best. This deficiency in
performance, not only for the Experimental Group but also
for the two Control Groups, cannot be attributed to the
185
subjects' lack of understanding of the demands of a
summarization task. As explained in Chapter V, prior to
conducting the study, the subjects expressed their awareness
of the requirements of a summarization task. Neither can
the writers' inability to reproduce a higher number of main
ideas be attributed to the 35-minute time limit for
completing the summarization task. Prior to data
collection, the subjects had been instructed to read and
become familiar with one of the three content-specific
articles.
The deficiency in performance, observed in all three of
the groups, could be attributed to a combination of two
problems. The subjects may have been insensitive to the
importance of textual ideas, and this problem may have been
aggravated by the partially descriptive nature of the texts
used in this study. The negative effect of descriptive
texts on the readers' recognition and recall of information
has been documented in the Ll and L2 literature (see
Carrell, 1987 for a discussion)1. The results regarding the
insensitivity to textual importance exhibited by poor
readers are supported by similar findings in Ll and L2
research (e.g., Johns & Mayes, 1992; Rinaudo, 1993;
Winograd, 1984). These findings suggest that the subjects
in this study would have benefited from instruction that
combined training in summarization, as operationalized in
186
this study, and training in text structure awareness,
including main idea identification.
Fourth, the Experimental Group exhibited a gain in the
percentage of adjacent main ideas included in the summaries,
as well as in the percentage of adjacent main ideas
connected via conjunctions. In addition, to connect two
adjacent main ideas, the three groups relied very heavily on
adversative and additive conjunctions. These findings are
partially consistent with those reported by Montafto-Harmon
(1988) regarding the additive structure of written text
produced by Mexican Spanish writers.
Fifth, prior to summarization instruction, the
Experimental subjects and Control Group 1 subjects in Health
Science— with some instances in Animal Biology and Cultural
Anthropology— exhibited a lack of objectivity, expressed
their opinions about the topic, or explicitly responded to
the source text. In contrast, the post-intervention results
showed that the subjects in the Experimental Group
maintained a more objective tone, opened their summaries
with an explication of the pragmatic condition of the task,
and tended to produce more macropropositions that presented
the gist of the article. However, this trend toward more
macropropositions was apparent only in the summaries written
by the subjects in Health Science and Animal Biology. The
Experimental subjects in these two content areas may have
187
benefited from the strong collaboration between the content
and the ESL instructors, both of whom emphasized the * ■
importance of reading for the overall message of the author
of the text.
Sixth, summarization-strategy instruction resulted in
the frequent use of appropriate reproduction strategies,
such as paraphrasing, and in the reduction of inappropriate
reproduction strategies such as copying and nearly copying.
In this respect, the performance of the Experimental
subjects was superior to that of the "able" subjects. The
Experimental subjects not only paraphrased more often than
did Control Group 2 subjects, but they also exhibited a
higher paraphrasing gain than did Control Group 2 subjects.
In contrast, Control Group 1 subjects not only increased the
number of copies, but they also copied more and paraphrased
less than did the Experimental subjects.
Although few instances of quotations were observed for
the three groups, more proficient summary writers integrated
their quotations into the summaries by using discourse
markers such as according to. Instead, summary writers with
low holistic scores failed to integrate the quotations into
the summaries.
Seventh, prior to summarization instruction, the
Experimental subjects employed local summarization
strategies when they reproduced individual idea units and
188
sentences. Following instruction, the Experimental subjects
were equally unable to apply global summarization
strategies. The Experimental subjects read at the paragraph
level, as indicated by the number of combinations across
sentences within paragraphs, but seemed to be unable to go
beyond the paragraph and abstract global meaning. In this
respect, the Experimental subjects performed as
"underprepared" readers. The fact that the post
intervention results showed that Control Group 1 subjects in
Animal Biology used global strategies was rather surprising.
However, the cognitive load placed on the Experimental
subjects is a likely explanation for this result. At the
time of completing the post-intervention task, the
Experimental subjects may have had a difficult time
controlling different aspects of the reading-to-write task.
Specifically, the Experimental subjects may have had limited
"cognitive space" (Kirkland & Saunders, 1991, p. 108) to
produce a higher number of main ideas than the subjects in
the other two groups, to combine them across paragraphs, and
to focus on issues related to grammatical accuracy.
The inability to use global reading strategies was also
exhibited by Control Group 2 subjects. In this case, the
"able” readers did not function as such. The Experimental
subjects, Control Group 1 subjects in Health Science and
Cultural Anthropology, and Control Group 2 subjects seemed
189
to be unable to go beyond the paragraph level and abstract
the global meaning of the articles.
Eighth, the summarization efficiency mean ratio
corresponding to the Experimental summaries with a sum-score
of 9 or above improved to a point that it was three times
higher than that achieved by Control Group 2 summaries in
the post-intervention task. This means that the summaries
produced by the Experimental subjects contained unimportant
information or distortions less often than did the summaries
produced by the "underprepared" subjects in Control Group 1
and the "able" subjects in Control Group 2. This higher
efficiency ratio combined with the higher percentage of
adjacent main ideas produced by the Experimental Group
subjects are indicative of the concise nature of the post
intervention summaries produced by the subjects in the
Experimental Group.
Conclusions
The overall results of this study underscore the
effectiveness of explicit, content-based instruction in
summary writing with "underprepared" L2 students. As
previously indicated, following instruction, the holistic
performance of the Experimental Group, consisting of
"underprepared" subjects, was significantly superior to that
of Control Group 2, consisting of "able" students.
190
It is apparent that pre-baccalaureate reading and
writing courses and the typical freshman English composition
class are not preparing language minority students to
complete such an essential academic literacy task as summary
writing successfully. This argument is supported directly
by the results of this study and indirectly by the large
numbers of language minority students who fail the CSU-
mandated Graduation Writing Assessment Requirement (GWAR).
Of particular concern was the performance of the "able"
subjects following instruction, which in many aspects, was
inferior to that of the "underprepared" subjects in the
Experimental Group. At the time of this study, the "able1 1
subjects were far more experienced than the subjects in the
other two groups. The superior performance of the
Experimental subjects is an indication that the mainstream
freshman English composition class in which the "able" L2
subjects participated may not have met their academic
literacy development needs. A combination of factors such
as the instructors' lack of preparation to deal with the
language problems of L2 students, and the nature of the
topics selected for reading and writing, which are often
inaccessible to L2 students who are not full participants of
the mainstream culture, may have been counterproductive for
the academic preparation of these L2 students. Furthermore,
a literature-based curriculum, characteristic of many
191
English composition courses, may have contributed to the
poor performance of the "able" subjects.
The success of the intervention can be attributed to
the features of the instructional program. First,
instruction in summary writing in this study was grounded in
the disciplines corresponding to the content courses in
which the students were enrolled. The journals from which
the readings were obtained were discipline-specific and were
selected by the content-area instructors prior to the
beginning of the quarter. Additionally, strategy
instruction was explicit and included models, peer
assistance, gradual transition from teacher- to student-
control, and opportunities for independent strategy use
(Crabe, 1993; Palincsar & David, 1991; Palincsar & Klerk,
1991). The instruction that the "underprepared" language
minority students received was specifically targeted to the
students' needs, unlike freshman English composition
classes, which often do not address the specific language
needs of language minority students and may place together
L2 and Ll students, the content-based adjunct classes in
which the students participated were comprised of language
minority students who shared similar backgrounds and needs.
All of the students were native speakers of Spanish and most
of them had been born in the U.S..or had lived in the
country for most of their lives. Host of the students had
192
participated in a number of L2 programs in the U.S. but,
upon entrance to the CSU system, were still "underprepared'1
to face the academic literacy demands of the general
education curriculum and, as such, were required to enroll
in pre-baccalaureate reading/writing courses.
Because most writing assignments in content classes
(for a description of writing requirements, see Carson,
Chase, & Gibson, 1992; Carson, et al., 1992; Horowitz, 1986;
Keller-Cohen & Wolfe, 1987; Ostler, 1980; Rusikoff, 1994)
involve reading and, at a minimum, preparing summaries of
some kind, "underprepared" language minority students from a
Spanish-speaking background enrolled in pre-baccalaureate
and freshman English composition courses will benefit from
instruction that includes reading-to-write tasks anchored in
the disciplines represented in the general education
courses. These types of tasks will prepare students for the
real demands of the real curriculum.
The success of explicit instruction in summary writing,
however, did not extend to two specific areas. First,
instruction did not result in an improvement in the
"underprepared" L2 students' ability to apply global reading
strategies in summarizing written text. These findings,
consistent with those in Ll and L2 research, are indicative
of the fragmented nature of the reading strategies employed
by the L2 students. Although instances of global strategy
193
use were identified in the summary protocols. Experimental
subjects and "able” subjects applied local reading
strategies more often and more consistently.
The other area in which explicit strategy instruction
was not successful involved the identification of a high
number of main ideas. Although the Experimental subjects
became more sensitive to textual importance, the mean number
of main ideas included in the summaries was low. As
indicated in this chapter, the Experimental subjects would
have benefited from a longer intervention that included both
summarization-strategy instruction and formal discourse
structure training. In this way, the subjects could have
developed a better understanding of how to identify main
ideas in complex texts.
This study did not address the subjects' grammatical
problems in the eight hours of summarization-strategy
training. A longer intervention could have included a
grammar focus, which could ultimately have resulted in
enhanced reader-based summaries. This grammatical focus
could have included a number of activities aimed at helping
raise the students' consciousness about their grammatical
problems.
194
Recommendations
Recommendations for Classroom Pedagogy
The results of this study suggest seven recommendations
for classroom pedagogy. First, L2 students from a Spanish
speaking background can greatly benefit from explicit
strategy instruction in summary writing. Explicit strategy
instruction, as operationalized in this study, includes
outlining the summarization strategies and developing
activities that explicitly model the strategies to be used.
Classroom activities should gradually move from teacher- to
student-controlled, and should provide students with
multiple opportunities to employ the strategies in a
cooperative setting.
Second, content-based L2 courses lend themselves to the
use of authentic materials grounded in different academic
content areas. Content-based L2 courses provide students
with opportunities to develop and accumulate background
knowledge on specific topics. In teaching summary writing,
much as in other cognitively demanding reading-to-write
tasks, the selection of appropriate materials is critical.
In the case of summarization instruction, particular
attention should be paid to the students' level of L2
proficiency and to the cognitive demands of the readings
selected.
195
Third, pre-baccalaureate reading and writing courses
and freshman composition classes should prepare students to
meet the demands of academic courses. Therefore, academic
tasks that require the integration of reading and writing
should be the primary focus of instruction. Because a great
variety of academic literacy tasks require summarization
strategies, instruction in summary writing is critical.
This recommendation is particularly important in light of
the fact that the holistic performance of the "able"
subjects in Control Group 2, eligible to take the CSU-
mandated GWAR, was inferior to the performance of the
"underprepared" subjects in the Experimental Group.
Fourth, instructional programs designed to help
language minority students make a transition from secondary
to tertiary study should be grounded in authentic academic
requirements. Programs such as the California State
University Summer Bridge Program and the Freshman Summer
Program at the University of California, Los Angeles, as
well as pre-baccalaureate reading and writing courses should
involve content-area reading-to-write tasks, including
summary writing. Explicit strategy instruction in these
types of preparatory programs would provide students with an
awareness of how to manage the cognitive load that difficult
content imposes on students.
196
Fifth, content-area instructors in general education
courses can play a critical role in preparing language
minority students to face the literacy demands of content
courses. Although the role of the content instructors was
not a major focus of this study, it was observed that the
content instructors exerted great influence on the quality
of the summaries produced by the Experimental subjects.
Remember that macropositions were identified only for those
Experimental subjects whose content instructors emphasized
the importance of reading and understanding the overall
message of the author of the text.
Sixth, the results of this study support the idea that
NSS students enrolled in high school English and content
classes could benefit greatly from explicit summarization
instruction. This preparation would help language minority
students make an easier transition into the demands of the
general education curriculum in university-level programs.
Seventh, this study was limited to native Spanish
speaking students. However, explicit instruction in summary
writing should not be confined to any single linguistic
group. "Underprepared" students from language backgrounds
other than Spanish should also benefit from explicit
content-based instruction in summary writing. In addition,
summarization instruction could be enhanced by combining
summarization instruction, as operationalized in this study,
197
and text structure training. For example, summarization
instruction in a problem-solution structure would involve
teaching students how to recognize a problem-solution
structure in a given text, completing a visual
representation with information from the text, and writing a
summary with information included in the visual
representation (see Armbuster, Anderson, & Ostertag, 1989
for step-by-step procedures).
Recommendations for Further Research
This study is part of a growing body of literature on
summary writing and l>2 students. However, this
investigation was limited in instructional and learner
variables. Future research should extend these areas. For
example, comparing the effect of text structure training to
that of summarization instruction, as operationalized in
this study, would be worthwhile. As indicated elsewhere in
this chapter, the effect of instruction on the students'
ability to identify main ideas was significant. However,
from this study, it is not clear whether text structure
training would have helped the Experimental subjects
identify an even greater number of main ideas. Further
research in this area would help understand this issue.
Also related to text structure is the question of
whether the findings of this investigation would hold across
198
readings with text structures different from those in this
study. Given that the partially descriptive nature of the
readings in this study was thought to affect the students'
ability to recall important information, it is possible to
assume that problem-solution texts would facilitate the
recall of high-level ideas.
Another variable that needs to be investigated is the
effect of instruction on the L2 students' ability to
transfer strategies over time, to new tasks, and to other
content areas. A longitudinal study would contribute to the
understanding of whether L2 students can apply the
strategies learned to various instructional contexts.
Further research could involve a number of learner
variables not addressed in this investigation. For example,
at this time, it is not known whether students from language
backgrounds different from Spanish would benefit from the
instructional strategies taught in this study. An
investigation in this area would help determine the
applicability of summarization-strategy training, as
operationalized in this study, to other language minority
groups.
The results of this investigation showed that advanced
L2 students can benefit from summarization instruction. It
still remains to be seen whether summarization instruction,
as operationalized in this study, would provide
199
instructional benefits to students with lower levels of L2
proficiency.
This study involved students considered to be
"underprepared" L2 readers. Further research could focus on
whether L2 students with differing reading skills would
benefit from summarization instruction. Because a number of
researchers found that reading ability had a significant
effect on the students' ability to use information from
source texts (e.g., Kennedy, 1985; Spivey, 1983, Winograd,
1984), it could be speculated that the effect of
summarization instruction might differ for students with
different reading skills.
Finally, several studies have shown that the ability to
process information transfers across languages (Cumming,
1989; Endres-Niggemeyer, Waumans, & Yamashita, 1991; Sarig,
1993). Given this line of research, an interesting issue to
be considered would involve measuring the students' ability
to process written text both in Ll and in L2, prior to and
following summarization instruction. This would provide
researchers with valuable information regarding whether
summarization instruction in L2 affects the students'
strategies when processing not only L2 but also Ll texts.
200
Motes
1. Prior to this investigation, the researcher knew that
the partially descriptive nature of the texts used would
constrain the readers' ability to recognize a number of main
ideas. However, in order to compare the performance of the
subjects in three content areas, it was critical to find
comparable articles both in length and text structure. This
concern guided a very extensive library search that resulted
in the identification of the articles used in this
investigation.
201
BIBLIOGRAPHY
Academic Program Improvement Workgroup on Support for
Underprepared Students. (1994). Report to the senior
vice chancellor, academic affairs. Long Beach, CA:
Office of the Chancellor. The California State
University.
Ackerman, J. M. (1989). Reading and writing in the
academy: A comparison of two disciplines. Unpublished
doctoral dissertation, Carnegie Mellon University.
Ackerman, J. M. (1990). Translating context into action.
In L. Flower, V. Stein, J. Ackerman, M. J. Kantz, K.
McCormick, & W. C. Peck (Eds.), Reading to write:
Exploring a cognitive & social process (pp. 173-193).
New York, NY: Oxford University Press.
Adamson, H. D. (1990). ESL students' use of academic
skills in content courses. English for Specific
Purposes, 9, 67-87.
Adamson, H. D. (1991). Academic competence. Journal of
Intensive English Studies, 5, 55-79.
Adamson, H. D. (1993). Academic competence: Theory and
classroom practice: Preparing ESL students for content
courses. White Plains, NY: Longman Publishing Group.
Alderson, J. C., & Urquhart, A. H. (1984). ESP tests: The
problem with student background discipline. (ERIC
Document Reproduction Service No. ED 275 163)
Alexander, P. A., Kulikowich, J. M., & Schulze, S. K.
(1994). How subject-matter knowledge affects recall
and interest. American Educational Research
Association Journal, 31(2), 313-337.
Anderson, J. R. (1983). The architecture of cognition,
Cambridge, MA: Harvard University Press.
Armbuster, B. B. (1991). Framing: A technique for
improving learning from science texts. In C. Santa &
D. Alvermann (Eds.), Science learning: Processes and
applications (pp. 104-113). Newark, DE: International
Reading Association.
202
Armbuster, B. B.r & Anderson, T. H. (1980). The effect of
mapping on the free recall of expository text, (Tech.
Rep. No 160). Urbana-Champaign: University of
Illinois, Center for the Study of Reading.
Armbuster, B. B., Anderson, T. H., & Ostertag, J. (1987).
Does text structure/summarization instruction
facilitate learning from expository text? Reading
Research Quarterly, 22(3), 331-346.
Armbuster, B. B., & Brown, A. L. (1984). Learning from
reading: The role of metacognition. In R. C.
Anderson, J. Osborn, & R. J. Tierney (Eds.), Learning
to read in American schools: Basal readers and content
texts (pp. 273-281). Hillsdale, NJ: Lawrence Erlbaum
Associates, Inc.
Bartholomae, D. (1985). Inventing the university. In M.
Rose (Ed.), When a writer can't write; Studies in
writer*s block and other composing problems (pp. 134-
165). New Vork: Guilford Press.
Bartlett, B. J. (1978). Top-level structure as an
organizational strategy for recall of classroom text.
Unpublished doctoral dissertation, Arizona State
University.
Bartlett, F. C. (1932). Remembering: A study In
experimental and social psychology, Cambridge:
Cambridge University Press.
Barnett, J. E. (1984). Facilitating retention through
instruction about text structure. Journal of Reading
Behavior, 16(1), 1-13.
Basham, C., S. (1986). Summary writing: A study in
textual and contextual constraints. Unpublished
doctoral dissertation, The University of Michigan.
Basham, C., Ray, R., & Whalley, E. (1993). CroBS-cultural
perspectives on task representation in reading to
write. In J. Carson & I. Leki (Eds.), Reading in the
composition classroom: Second language perspectives
(pp. 161-182). Boston, MA: Heinle & Heinle
Publishers.
Bensoussan, M., & Kreindler, I. (1990). Improving advanced
reading comprehension in a foreign language: Summaries
vs. short-answer questions. Journal of Research in
Reading, 13(1), 55-68.
203
Benesch, S. (1988). Linking content and language teachers:
Collaboration across the curriculum. In S. Benesch
(Ed.), Ending remediation: Linking ESL and content in
higher education (pp. 57-65).
Bereiter, C., & Bird, H. (1985). Use of thinking aloud in
identification and teaching of reading comprehension
strategies. Cognition and Instruction, 2, 131-156.
Bereiter, C., & Scardamalia, H. (1987). The psychology of
written composition. Hillsdale, NJ: Lawrence Erlbaum
Associates, Inc.
Berkowitz, s. J., (1986). Effects of instruction in text
organization on sixth-grade students' memory for
expository reading. Reading Research Quarterly, 21(2),
161-178.
Bower, B. (1988). Murder in good company. Science News,
133(5), 90-91.
Bower, B. (1992). Here comes the sun. Science News,
142(1), 142-143.
Boyle, 0. F., & Peregoy, S. F. (1990). The effects of
cognitive mapping on students' learning from college
texts. Journal of College Reading and Learning, 23(2),
14-22.
Braine, G. (1995). Writing in the natural sciences and
engineering. In D Belcher & G. Braine (Eds.), Academic
writing in a second language: Essays on research and
pedagogy (pp. 113-134). Norwood, NJ: Ablex Publishing
Corporation.
Bridgeman, B., & Carlson, S. B. (1984). Survey of academic
writing tasks. Written Communication, 1, 247-280.
Brinton, D. M., & Snow, M. A. (1988). Content-based
language instruction: Investigating the effect of the
adjunct model. TESOL Quarterly, 22(4), 553-574.
Brinton, D. M., Snow, M. A., 6 Wesche, M. B. (1989).
Content-based second language instruction. New York,
NY: Newbury House Publishers.
Britton, J., Burguess, T., Martin, N., McLeod, A., & Rosen,
H. (1978). The development of writing abilities.
Urbana, IL: National Council of Teachers of English.
204
Brooks, L. w., & Danserau, D. F. (1983). Effects of
structural schema training and text organization on
expository prose processing. Journal of Educational
Psychology, 75(6), 811-820.
Brown, A. L, Campione, J. C., 6 Day, J. D. (1981).
Learning to learn: On training students to learn from
texts, Educational Researcher, 10(2), 14-21.
Brown, A. L., & Day, J. D. (1983). Macrorules for
summarizing texts: The development of expertise.
Journal of Verbal Learning and Verbal Behavior, 22(1),
1-14.
Brown, A. L., Day, J. D., 6 Jones, R. s. (1983). The
development of plans for summarizing texts, child
Development, 54, 968-979.
Brown, A. L., & Palincsar, A. S. (1989). Guided,
cooperative learning and individual knowledge
acquisition. In L. B. Resnick (Ed.), Knowing, learning
and instruction: Essays in honor of Robert Glaser (pp.
393-451). Hillsdale, NJ: Lawrence Erlbaum Associates,
Inc.
Brown, A. L., Palincsar, A. S., & Armbuster, B. B. (1994).
Instructing comprehension-fostering activities in
interactive learning situations. In R. B. Ruddell, M.
R. Ruddell, & H. Singer (Eds.), Theoretical models and
processes of reading (4th ed., pp. 757-787). Newark,
DE: International Reading Association.
Campbell, C, C. (1987). Writing with others' words. The
use of information from a background reading text in
the writing of native and nonnative university
composition students. Unpublished doctoral
dissertation, University of California, Los Angeles.
Campbell, C. C. (1990). Writing with others' words: Using
background reading text in academic compositions. In
B. Kroll (Ed.), Second language writing: Research
insights for the classroom (pp. 211-230). Cambridge:
Cambridge University Press.
Campione, J. C., Brown, A. L., & Ferrara, R. A. (1982).
Mental retardation and intelligence. In R. Sternberg
(Ed.), Handbook of human intelligence (pp. 231-235).
Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.
205
Carrell, P. L. (1984a)* Evidence of a formal schema in
second language comprehension. Language Learning,
34(2), 87-112*
Carrel1, P. L. (1984b). The effects of rhetorical
organization on ESL readers. TESOL Quarterly, 18(3),
441-469.
Carrell, P. L. (1985). Facilitating ESL reading by
teaching text structure. TESOL Quarterly, 19(4), 727-
752.
Carrell, P. L. (1987). Text as interaction: Some
implications of text analysis and reading research for
ESL composition. In U. Connor & R. B. Kaplan (Eds.),
Writing across languages: Analysis of L2 text (pp. 47-
56). Reading, MA: Addison-Wesley Publishing Company.
Carrell, P. L. (1992). - Awareness of text structure:
Effects on recall. Langruage Learning, 42(1), 1-20.
Carrell, P. L., & Eisterhold, J. C. (1983). Schema theory
and ESL reading pedagogy. TESOL Quarterly, 17(4), 553-
573.
Carrell, P. L., & Eisterhold, J. C. (1988). Schema theory
and ESL reading pedagogy. In P. L. Carrell, J. Devine,
& D. E. Eskey (Eds.), Interactive approaches to second
language reading (pp. 73-92). Cambridge: Cambridge
University Press.
Carrell, P. L., Pharis, B. 6., & Liberto. J. C. (1989).
Metacognitive strategy training for ESL reading. TESOL
Quarterly, 23(4), 647-673.
Carson, J. G. (1993). Reading for writing: Cognitive
perspectives. In J. G. Carson & I. Leki (Eds.),
Reading in the composition classroom: Second language
perspectives (pp. 85-104). Boston, MA: Heinle and
Heinle Publishers.
Carson, J. G., Chase, N. D., & Gibson, S. U. (1992).
Literacy analyses of high school and university
courses: Summary descriptions of selected courses.
Atlanta, GA: Center for the Study of Adult Literacy.
Georgia State University.
Carson, J. G., Chase, N. D, Gibson, S. U, & Hargrove, M.
(1992). Literacy demands of the undergraduate
curriculum. Reading Research and Instruction, 31 (4),
25-50.
206
Casazza, M. E. (1993). Using a model of direct instruction
to teach summary writing in a college reading class.
Journal of Reading, 37(3), 202*208.
Caverly, D., & Orlando, V. (1991). Textbook study
strategies. In R. Flippo 6 D. Caverly (Eds.), Teaching
reading and study strategies at the college level (pp.
86*165). Newark, DE: International Reading
Association.
Chamot, A. U., Dale, M., O'Malley, J. M., & Spanos, G. A.
(1992). Learning and problem solving strategies of ESL
students. Bilingual Research Journal, 16(3-4), 1-34.
Charry, M. B. (1987). Underprepared community college
students: A pilot experiment. Research and Teaching
Developmental Education, 3(1), 22-30.
Clark, R. (1990). A cognitive theory of instructional
method. Paper presented at the annual meeting of the
American Educational Research Association. Boston, MA.
Coady, J. (1979). A psycholinguistic model of the ESL
reader. In R. Mackay, R. Barkman, & R. Jordan (Eds.),
Reading in a second language (pp. 5-12). Rowley, MA:
Newbury House.
Cohen, A. (1993). The role of instructions in testing
summarizing ability. In D. Douglas & c. Chapelle
(Eds.), A new decade of language testing research:
Selected papers from the 1990 language testing research
colloquium (pp. 132-160). Alexandria, VA: Teachers of
English to Speakers of Other Language.
Commander, N. E., 6 Smith, B. D. (1995). Developing
adjunct reading and learning courses that work.
Journal of Reading 38(5), 352-360.
Committee on Educational Policy. (1995). Precollegiate
instruction in the CSU. Long Beach, CA: Office of the
Chancellor. The California State University.
Connor, U. M. (1984). Recall of text: Differences between
first and second language readers. TESOL Quarterly,
18(2), 239-256.
207
Connor, U. M., 6 Fanner, M. (1990). The teaching of
topical structure analysis as a revision strategy for
ESL writers. In B. Kroll (Ed.), Second language
writing: Research Insights for the classroom (pp. 126-
139). Cambridge: Cambridge University Press.
Connor, U. M., & HcCagg, P. (1987). A contrastive study of
English expository prose paraphrases. In U. Connor &
R. B. Kaplan (Eds.), (Writing across languages:
Analysis of L2 text (pp. 73-86). Reading, MA:
Addison-Wesley Publishing Company.
Connor, U. M., & Kramer, M. G. (1995). Writing from
sources: Case studies of graduate students in business
management. In D Belcher & G. Braine (Eds.), Academic
writing in a second language: Essays on research and
pedagogy (pp. 155-182). Norwood, NJ: Ablex Publishing
Corporation.
Collier, V. (1987). Age and rate of acquisition of second
language for academic purposes. TESOL Quarterly,
21(4), 617-641.
Collier, V. (1989). How long? A synthesis of research on
academic achievement in a second language. TESOL
Quarterly, 23(4), 509-531.
cordero-Ponce, W. L. (1994). Facilitating L2 reading
comprehension through summarization instruction. Paper
presented at the annual conference of the American
Association of Applied Linguistics, Baltimore, MD.
Crandall, J. (1992). Content-centered learning in the
United States. Annual Review of Applied Linguistics,
13, 111-126.
Cronbach, L. J., Gleser, G. C., Nanda, H., & Rajaratnam, N.
(1972). The dependability of behavioral measurements:
Theory of generalizability for scores and profiles.
New York: John Wiley.
dimming, A. (1989a). Reading and summarizing challenging
texts in first and second languages. (ERIC Document
Reproduction Service No ED 306 773)
dimming, A. (1989b). Writing expertise and second language
proficiency. Language Learning, 39(1), 81-141.
208
Cummins, J. (1981). The role of primary language
development in promoting educational success for
language minority students. In Schooling and language
minority students: A theoretical framework (pp. 3*49).
Los Angeles, CA: Evaluation, Dissemination and
Assessment Center, California State University, Los
Angeles.
Cummins, J. (1989). Empowering minority students.
Sacramento, CA: California Association for Bilingual
Education.
Davey, B., & McBride, S. (1986). Effects of question*
generation training on reading comprehension. Journal
of Educational Psychology, 70(2), 256-262.
Day, J. D. (1980). Teaching summarization skills: A
comparison of training methods. Unpublished doctoral
dissertation, University of Illinois, Urbana.
Derry, S. J., & Murphy, D. A. (1986). Designing systems
that train ability: Prom theory to practice. Review
of Educational Research, 56(1), 1-39.
Devine, J. (1988). The relationship between general
language competence and second language reading
proficiency: Implications for teaching. In P. L.
Carrel1, J. Devine, & D. E. Eskey (Eds.), Xnteractive
approaches to second language reading (pp. 260-277).
Cambridge: Cambridge University Press.
Doughty, C. (1991). Second language instruction does make
a difference: Evidence from an empirical study of
relativization. Studies in Second Language
Acquisition, 13, 431-469.
Dubin, F. (1986). Dealing with texts. In F. Dubin, D. E.
Eskey, & W. Grabe (Eds.), Teaching second language
reading for academic purposes (pp. 127-160). Reading,
MA: Addison-Wesley Publishing Company.
Durst, R. (1984). The development of analytic writing. In
A. N. Applebee (Ed.), Contexts for learning to write.
Norwood, NJ: Ablex Publishing Corporation.
Duffy, C. B. (1995). Adjunct-model content based language
instruction: Effects on language and concept
development. Paper presented at the annual conference
of the American Association of Applied Linguistics, '
Long Beach, CA.
209
Edelsky, C. (1982). Writing in a bilingual program. The
relation of Ll and L2 teste. TBSOL Quarterly, 16(2),
211-228.
Edelsky, C. (1983). Segmentation and punctuation:
Developmental data from young writers in a bilingual
program. Research In the Teaching of English, 17(2),
135-156.
Edelsky, C. (1993). Habia una vez: Writing in a bilingual
program (2nd ed.). Norwood, NJ: Ablex Publishing
Corporation.
Ellis, R. (1990). Instructed second language acquisition.
Oxford: Blackwell.
Ellis, R. (1993). The structural syllabus and second
language acquisition. TBSOL Quarterly, 27(1), 91-111.
Endres-Niggemeyer, B., Waumans, W., 6 Yamashita, H. (1991).
Modeling summary writing by introspection: A small-
scale demonstrative study. TEXT, 11(4), 523-552
English as a Second Language Workgroup. (1988). Report.
Long Beach, CA: Office of the Chancellor. The
California State University.
Eskey, D. E. (1988). Holding in the bottom: An
interactive approach to the language problems of second
language readers. In P. L. Carrell, J. Devine, & D. E.
Eskey (Eds.), Interactive approaches to second language
reading (pp. 93-100). Cambridge: Cambridge University
Press.
Eskey, D. E. (1992). Syllabus design in content-based
instruction. The CATESOL Journal, 5(1), 11-23.
Eskey, D. E., & Grabe, W. (1988). Interactive models for
second language reading: Perspectives on instruction.
In P. L. Carrell, J. Devine, & D. E. Eskey (Eds.),
Interactive approaches to second language reading (pp.
223-238). Cambridge: Cambridge University Press.
Fackelmann, K. A. (1991). The safer sex: Probing a
cardiac gender gap. Science Hews, 139(3), 139-140.
Ferris, D. (1992). cross cultural variation in SSL
students1 responses to an essay prompt. Unpublished
manuscript. University of Southern California.
210
Ferris, D. (1995). Student reactions to teacher response
in multiple draft composition classrooms. TSSOL
Quarterly, 29(1), 33-53.
Flower, L. (1979). Writer-based prose: A cognitive basis
for problems in writing. College English, 41(1), 19-
37.
Flower, L. (1987). Interpretive acts: Cognition and the
construction of discourse. Poetics, 16, 109-130.
Flower, L. (1989). Cognition, context, and theory
building. College Composition and Communication,
40(3), 282-311.
Flower, L. (1990). The role of task representation in
reading-to-write. In L. Flower, V. Stein, J. Ackerman,
M. J. Kantz, K. McCormick, & W. C. Peck (Eds.), Reading
to write: Exploring a cognitive & social process (pp.
35-75). New York, NY: Oxford University Press.
Flower, L., & Hayes, J. (1980). The cognition of
discovery: Defining a rhetorical problem. College
Composition and Communication, 31(1), 21-32.
Fry, E. (1977). Fry*s readability graph: Clarifications,
validity, and extension to level 17. Journal of
Reading, 21(3), 242-253.
Gaies, S. J. (1991). ESL students in academic courses:
Forging a link. College ESL, 1(1),30-36.
Garner, R. (1982). Efficient text summarization: Costs
and benefits. Journal of Educational Research, 75(5),
275-279.
Garner, R. (1985). Text summarization deficiencies among
older students: Awareness or production ability?
American Educational Research Journal, 22(4), 549-560.
Garner, R., Belcher, V., Winfield, E., & Smith, T. (1985).
Multiple measures of text summarizing proficiency:
What can fifth-grade students do? Research in the
Teaching of English, 19(2), 140-153.
Garner, R., 6 McCaleb, J. L. (1985). Effects of text
manipulation on quality of written summaries.
Contemporary Educational Psychology, 10(2), 139-149.
211
Gee, Y. (1992). How can ESL and content teachers work
effectively together in adjunct courses? The CATESOL
Journal, 5(1), 85-92.
Genesee, F. (1987). Learning through two languages:
Studies of Immersion and bilingual education.
Cambridge, MA: Newbury House.
Geva, E. (1983). Facilitating reading comprehension
through flowcharting. Reading Research Quarterly,
18(4), 384-405.
Geva, E. (1992). The role of conjunctions in L2 text
comprehension. TESOL Quarterly, 26(A), 731-747.
Geva, E., & Ryan, E. B. (1985). Use of conjunctions in
expository texts by skilled and less skilled readers.
Journal of Reading Behavior, 17(4), 331-346.
Goldstein, L., Campbell, C., & Cummings, M. c. (1994).
Smiling through the turbulence: The flight attendant
syndrome and other issues of writing instructor status
in the adjunct model. The CATESOL Journal, 7(1), 19-
29.
Gordon, C. J. (1990). Contexts for expository text
structure use. Reading Research and Instruction,
29(2), 55-72.
Grabe, W. (1986). The transition from theory to practice
in teaching reading. In F. Dubin, D. E. Eskey, & W.
Grabe (Eds.), Teaching second language reading for
academic purposes (pp. 25-48). Reading, HA: Addison-
Wesley Publishing Company.
Grabe, W. (1993). Rethinking reading strategies. Paper
presented at the annual convention of Teachers of
English to Speakers of Other Languages, Atlanta, GA.
Grimes, J. E. (1975). The thread of discourse. The Hague:
Houton.
Guido, B., & Colwell, C. G. (1987). A'rationale for direct
instruction to teach summary writing following
expository text reading. Reading Research and
Instruction, 26(2), 89-98.
Guri-Rozenblit, S. (1989). Effects of a tree diagram on
students' comprehension of main ideas in an expository
text with multiple themes. Reading Research Quarterly,
24(2), 236-247.
212
Halliday, H. A. K., & Hasan, R. (1976). Cohesion In
English, London: Longman Publishing Group.
Hare, V., & Borchardt, K. (1984). Direct instruction of
summarising skills. Reading Research Quarterly, 20(1),
62-78.
Harley, B, (1994). Appealing to consciousness in the L2
classroom. Applied Linguistics, 11, 57-68.
Hern&ndez, J. S. (1991). Assisted performance in reading
comprehension strategies with non-English proficient
students. The Journal of Educational Issues of
Language Minority students, 8, 91-112.
Hidi, S. (1985). Variables that affect how children
summarize school texts and the amount they learn during
this activity. Paper presented at the annual meeting
of the American Educational Research Association,
Chicago, IL.
Hidi, S., & Anderson, V. (1986). Producing written
summaries: Task demand, cognitive operations, and
implications for instruction. Review in Educational
Research, 56(4), 473-493.
Hinds, J. (1983). Linguistic and written discourse in
English and Japanese: A contrastive study (1978-1982).
Annual Review of Applied Linguistics, 3. Rowley, HA:
Newbury House.
Hinds, J. (1987). Reader versus writer responsibility: A
new typology. In U. Connor & R. B. Kaplan (Eds.),
Writing across languages; Analysis of L2 text (pp.
142-152). Reading, HA: Addison-Wesley Publishing
Company.
Holmes, J., & Ramos, R. G. (1993). Study summaries as an
evaluation instrument: Questions of validity. English
for Specific Purposes, 12(1), 83-94.
Horowitz, D. (1986). What professors actually require:
Academic tasks for the ESL classroom. TESOL Quarterly,
20(3), 445-462.
Hudson, T. (1988). The effects of induced schemata on the
"short circuit" in L2 reading: Non-decoding factors in
L2 reading performance. In P. Carrell, J. Devine, & D.
E. Eskey (Eds.), Interactive approaches to second
language reading (pp. 183-205). Cambridge: Cambridge
University Press.
213
Hudson, T. (1991). A content-comprehension approach to
reading in science and technology. TESOL Quarterly,
25(1), 77-104.
Hulstijn, J. H., & Schmidt, R. (1994). Association
Internationale de Llngulstique AppliquSe Review, 11.
Hulstijn, J* H., & Schmidt, R. (1994). Guest editors'
introduction* Association Internationale de
Llngulstique Appliqu6e Review, 11, 5-10.
Johns, A. H. (1985). Summary protocols of "under-prepared"
and "adept" university students: Replications and
distortions of the original. Language Learning, 35(4),
495-517.
Johns, A. H. (1988a). The discourse communities dilemma:
Identifying transferable skills for the academic
milieu. English for Specific Purposes, 7(1), 55-60.
Johns, A. M. (1988b). Reading for summarising: An
approach to text orientation and processing. Reading
in a Foreign Language, 4(2), 79-90.
Johns, A. M. (1991a). Faculty assessment of ESL student
literacy skills: Implications for writing assessment.
In L. Hamp-Lyons (Eds.), Assessing second language
writing In academic contexts (pp. 167-179). Norwood,
NJ: Ablex Publishing Corporation.
Johns, A. H. (1991b). English for specific purposes (ESP):
Its history and contributions. In M. Celce-Murcia
(Ed.), Teaching English as a second or foreign language
(pp. 67-78). New York: Newbury House.
Johns, A. M. (1991c). Interpreting an English competency
examination: The frustrations of an ESL science
student. Written Communication, 8(3), 379-401.
Johns, A. H. (1992). What is the relationship between
content-based instruction and English for specific
purposes? The CATESOL Journal, 5(1), 71-75.
Johns, A. H., & Dudley-Evans, T. (1991). ESP:
International in scope, specific in purpose. TESOL
Quarterly, 25(2), 297-314.
Johns, A. H., & Hayes, P. (1990) An analysis of summary
protocols of university ESL students. Applied
Linguistics, 11, 253-271.
214
Johnson, H. (1983). What do you do if you can't tell the
whole story? The development of summarization skills.
In K. E. Nelson (Ed.), Children's language (Vol. 4, pp.
315-383). Hillsdale, NJ: Lawrence Erlbaum Associates,
Inc.
Johnson, R. E. (1970). Recall of prose as a function of
structural importance of the linguistic units. Journal
of Verbal Learning and Verbal Behavior, 9, 12-20.
Kantz, M. (1987). Synthesizing rhetorical stance:
Undergraduates doing a reading to write task.
Unpublished doctoral dissertation, Carnegie Mellon
University, Pittsburgh.
Kaplan, R. B. (1966). Cultural thought patterns in
intercultural education. Language Learning, 16, 1-20.
Kaplan, R. B. (1988). Contrastive rhetoric and second
language learning: Notes toward a theory of
contrastive rhetoric. In A. C. Purves (Ed.), Writing
across languages and cultures: Issues in contrastive
rhetoric (pp. 275-304). Newbury Park, CA: SAGE.
Kaplan, R. B., Cantor, S., Hagstrom, C., Kamhi-Stein, L. D.,
Sniotani, Y., & Zimmerman, C., B. (1994). On abstract
writing. TEXT, 14(3), 401-426.
Keller-Cohen, D., & Wolfe, A. (1987). Extending writing in
the College of Literature, Science and the Arts:
Report of a faculty survey. Unpublished manuscript,
University of Michigan, Ann Arbor.
Kennedy, M. L. (1985). The composing process of college
students writing from sources. Written Communication,
2(4), 434-456.
King, A. (1992). Comparison of self-questioning,
summarizing, and notetaking-review as strategies from
learning from lectures. American Educational Research
Journal, 29(2), 303-323.
Kinsella, K. (1992). How can we move from comprehensible
input to active learning strategies in content-based
instruction? The CATESOL Journal, 5(1), 127-132.
Kintsch, W., & van Dijk, T. A. (1978). Toward a model of
text comprehension and production. Psychological
Review, 85(5), 363-394.
215
Kirkland, M. R., & Saunders, H. A. P. (1991). Maximizing
student performance in summary writing: Managing
cognitive load. TESOL Quarterly, 25(1), 105-121.
i.
Kozminsky, E., & Graetz, N. (1986). First vs second
language comprehension: Some evidence from text
summarizing. Journal of Research in Reading, 9(1), 3-
21.
Krashen, S. (1981). The case for narrow reading. TESOL
Newsletter, 15, 23.
Krashen, S. (1993). The power of reading: Insights from
the research. Englewood, CO: Libraries Unlimited,
Inc.
Krashen, S., & Biber, D. (1988). On course: Bilingual
education's success in California. Sacramento, CA:
California Association for Bilingual Education.
Kroll, B. (1977). Combining ideas in written and spoken
English: A look at subordination and coordination. In
E. 0. Keenan & T. L. Bennet (Eds.), Discourse across
time and space (pp. 69-108). Southern California
Occasional Papers in Linguistics, Vol. 5. Los Angeles:
University of Southern California.
Kucer, S. L. (1985). The making of meaning: Reading and
writing as parallel processes. Written Communication,
2(3), 317-336.
Lambert, W. E., & Tucker, G. R. (1972). Bilingual
education of children. Rowley, MA: Newbury House
Publishers.
Langer, J. A., & Applebee, A. N. (1987). Now writing
shapes thinking. Research Report No. 22. Urbana, IL:
National Council of Teachers of English.
Lautamatti, L. (1987). Observations on the development of
topic of simplified discourse. In U. Connor & R. B.
Kaplan (Eds.), Writing across languages: Analysis of
L2 text (pp. 87-114). Reading, MA: Addison-Wesley
Publishing Company.
Leki, I., & Carson, J. G. (1994). Students' perceptions of
EAP writing instruction and writing needs across the
disciplines. TESOL Quarterly, 28(1), 81-101.
216
Macias, R. F. (1993). Language and ethnic classification
of language minorities: Chicano and Latino students in
the 1990s. Hispanic Journal of Behavioral Sciences,
15(2), 230-257.
Maimon, E. O. (1979). Talking to strangers. College
Composition and Communication, 30(4), 364-369
Marcoulides, G. A. (1989). The application of
generalizability analysis to observational studies.
Quality & Quantity, 23, 115-127.
Master, P. (1992). What are some considerations for
teacher training in content-based instruction? The
CATESOL Journal, 5(1), 77-84.
Maxworthy, A. G., & Barry, A. (1992). Can middle school
students summarize? Reading Horizons, 32 (3), 191-198.
Meyer, B. J. F. (1975). Identification of the structure of
prose and its implications for the study of reading and
memory. Journal of Reading Behavior, 75(7), 7-47.
Meyer, B. J. F. (1977). Organization in prose and memory.
Research with application to reading comprehension. In
P. D. Pearson (Ed.), Reading: Theory, research, and
practice (pp. 214-220). Clemson, SC: National Reading
Conference.
Meyer, B. J. F. (1979). Organizational patterns in prose
and their use in reading. In M. L. Kamil & A. J. Moe
(Eds.), Reading research: Studies and applications
(pp. 109-117). Clemson, SC: National Reading
Conference.
Meyer, B. J. F. (1981). Prose analysis: Procedures,
purposes, and problems (Prose Learning Series,
Research Report No 11). Tempe: Arizona State
University. (ERIC Document Reproduction Service No ED
201 972)
Meyer, B. J. F. (1982). Reading research and the
composition teacher: College Composition and
Communication, 33(1), 37-49.
Meyer, B. J. F., Brandt, D. M., & Bluth, G. J. (1980). Use
of top-level structure in text: Key for comprehension
of ninth-grade students. Reading Research Quarterly,
16(1), 72-103.
217
Meyer, B. J. F., 6 Freedle, R. O. (1984). Effects of
discourse type on recall. American Educational
Research Journal, 21(1), 121-143.
Meyer, B. J. F., & McConkie, G. W. (1973). What is
recalled after hearing a passage? Journal of
Educational Psychology, 85(1), 109-117.
Mohan, B. (1986). Language and content. Reading, MA:
Addison-Wesley Publishing Company.
Montafio-Harmon, M. R. (1988). Discourse features in the
compositions of Mexican, Engllsh-as-a-second-language,
Mexican-American/Chicano, and Anglo high school
students: Considerations for the formulation of
educational policies. Unpublished doctoral
dissertation, University of Southern California.
Murray, D. E., Nichols, P. C., & Heisch, A. (1992).
Identifying the languages and cultures of our students.
In D. Murray (Ed.), Diversity as resource: Redefining
cultural literacy (pp. 63-83). Alexandria, VA:
Teachers of English to Speakers of Other Languages.
Mushakoji, S. (1991). The process of producing scholarly
author abstract: Discourse analysis. Proceedings for
Gesellschaft fur Angewandte Linguistik c.v.,
Unlversitat Mainz, (Vol. 26) zum 28. Ramenthema:
Wirtschaft Und Sprache.
Nielsen, H. B. (1995). Seductive texts with serious
intentions. Educational Researcher, 24(1), 4-12.
Nist, S., & Kirby, K. (1986). Teaching comprehension and
study strategies through modeling and thinking aloud.
Reading Research and Instruction, 25(4), 254-264.
Norusis, M. J. SPSS for windows: Base system user's guide,
Release 6.0. (1993). Chicago, IL: SPSS Inc.
O'Malley, J. M., & Chamot, A. U. (1990). Learning
strategies in second language acquisition. Cambridge:
Cambridge University Press.
Oshima, A., & Hogue, A. (1991). Writing academic English.
Menlo Park, CA: Addison-Wesley Publishing Company.
Ostler, S. (1980). A survey of academic needs for advanced
ESL. TESOL Quarterly, 14(4), 489-502.
218
Padrdn, Y. K. (1985). Utilizing cognitive reading
strategies to improve English reading comprehension of
Spanish-speaking bilingual students. Unpublished
doctoral dissertation, University of Houston.
Padrdn, Y. H. (1992). The effect of strategy instruction
on bilingual students' cognitive strategy use in
reading. Bilingual Research Journal, 16(3-4), 35-52.
Padrdn, Y. N., Knight, S. L., & Waxman, H. C. (1986).
Analyzing bilingual and monolingual students'
perceptions of their reading strategies. The Reading
Teacher, 39(5), 430-433.
Padr6n, Y. N., & Waxman, H. C. (1988). The effect of ESL
students' perceptions of their cognitive reading
strategies on reading achievement. TESOL Quarterly,
22(1), 146-150.
Palincsar, A. S., & Brown, A. L. (1984). Reciprocal
teaching of comprehension-fostering and comprehension-
monitoring activities. Cognition and Instruction,
1(2), 117-175.
Palincsar, A. S., & David, Y. (1991). Promoting literacy
through classroom discourse. In E. Hiebert (Ed.),
Literacy for a diverse society (pp. 122-140). New
York: Teachers College Press.
Palincsar, A. S., & Klerk, L. (1991). Dialogues promoting
reading comprehension. In B. Means, C. Chelemer & M.
Knapp (Eds.), Teaching advanced skills to at-risk
students (pp. 112-140). San Francisco, CA: Jossey-
Bass.
Paris, S. G., & Myers, M. (1981). Comprehension
monitoring, memory, and study strategies of good and
poor readers. Journal of Reading Behavior, 13, 5-22.
Patthey-Chavez, G. G. (1988). Writing opinions in high
school: A comparison of Anglo and Latino students*
texts. Paper presented at the annual conference of the
American Association of Applied Linguistics, New
Orleans, LA.
Pearson, P. D., & Camperell, K. (1994). Comprehension of
text structures. In R. B. Ruddell, M. R. Rudell, & H.
Singer (Eds.), Theoretical models and processes of
reading (4th ed., pp. 448-468). Newark, DE:
International Reading Association.
219
Perl, S. (1979). The composing processes of unskilled
college writers. Research in the Teaching of English,
13 (4), 364-382.
Pressley, M., Levin, J. R., & Ghatala, E. s. (1984).
Metamemory strategy memory in adults and children.
Journal of Verbal Learning and Verbal Behavior, 23,
270-288.
Rafoth, B. A. (1985). Audience adaptation in the essays of
proficient and nonproficient freshman writers.
Research in the Teaching of English, 19(3), 237-253.
Raphael, T. E., & Pearson, P. D. (1985). Increasing
student awareness of sources of information for
answering questions. American Educational Research
Journal, 22(2), 217-237.
Raymond, P. M. (1993). The effects of structure strategy
training on the recall of expository prose for
university students reading French as a second
language. The Modern Language Journal, 77(4), 445-458.
Ready, D., & Wesche, M. (1992). An evaluation of the
University of Ottawa's sheltered program: Language
teaching strategies that work. In R. Courchene, J.
Glidden, R. St. John, & C. Therien (Eds.),
Comprehension second language teaching, (pp. 389-405).
Ottawa: University of Ottawa Press.
Reid, J. (1990). Responding to different topic types: A
quantitative analysis from a contrastive rhetoric
perspective. In B. Kroll (Ed.), Second language
writing: Research insights for the classroom (pp. 191-
210). Cambridge: Cambridge University Press.
Ricciardi, P. (1984). Children’s selection of the critical
elements of prose passages. Paper presented at the
meeting of the American Psychological Association,
Toronto, Canada.
Rinaudo, M. C. (1993). Metacognicidn y estrategias de
aprendizaje [Metacognition and learning strategies].
Lectura y Vida, 14(3), 5-12.
Rinehart, S. D., & Thomas, K. F. (1993). Summarization
ability and text recall by novice studiers. Reading
Research and Instruction, 32(4), 24-32.
Rose, M. (1983). Remedial writing courses: A critique and
a proposal. College English, 45(2), 109-128.
220
Rosenshine, B., & Meister, C. (1994). Reciprocal teaching:
A review of the research* Review of Educational
Research, 64(4), 479-530.
Rumelhart, 0* E. (1977a). Toward an interactive model of
reading. In S. Dorniv (Ed.), Attention and performance
VI (pp. 573-606). Hillsdale, NJ: Lawrence Erlbaum
Associates, Inc.
Rumelhart, D. E. (1977b). Understanding and summarizing
brief stories. In D. Laberge & S. J. Samuels (Eds.),
Basic processes in reading: Perception and
comprehension (pp. 265-303). Hillsdale, NJ: Lawrence
Erlbaum Associates, Inc.
Rumelhart, D. E. (1980). Schemata: The building blocks of
cognition. In R. Spiro B. Bruce, & W. Weber (Eds.),
Theoretical issues in reading comprehension (pp. 33-
58). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.
Rumelhart, D. E., & Ortony, A. (1977). The representation
of knowledge in memory. In R. C. Anderson, R. J.
Spiro, & W. E. Montagna (Eds.), Schooling and the
acquisition of knowledge (pp. 99-135). Hillsdale, NJ:
Lawrence Erlbaum Associates, Inc.
Rusikoff, K. A. (1994). A comparison of selected writing
criteria used to evaluate non-native speakers of
English at a California State University. Unpublished
doctoral dissertation, University of Southern
California.
Sarig, G. (1987). High-level reading in the first and in
the foreign language: Some comparative process data.
In J. Devine, P. Carrell, & D. Eskey (Eds.), Research
in reading in English as a second language (pp. 106-
120). Washington, DC:: Teachers of English to
Speakers of Other Languages.
Sarig, G. (1993). Composing a study-summary: A
reading/writing encounter. In J. Carson & I. Leki
(Eds.), Reading in the Composition Classroom: Second
Language Perspectives (pp. 161-182). Boston, MA:
Heinle & Heinle Publishers.
Schmidt, R. (1990). The role of consciousness in second
language learning. Applied Linguistics, 11, 17-46.
221
Schmidt, R. (1992). Awareness and second language
acquisition. Annual Review of Applied Linguistics, 13,
206-226.
Sharwood Smith, M. (1981). Consciousness raising and the
second language learner. Applied Linguistics, 2, 159-
168.
Shavelson, R., Webb, N. M., & Rowley, G. L. (1989).
Generalizability theory. American Psychologist, 44(6),
922-932.
Sherrard, C. (1989). Teaching students to summarize:
Applying textlinguistics. System, 17(1), l-ll.
Shih, H. (1986). Content-based approaches to teaching
academic writing. TESOL Quarterly, 20(4), 617-648.
Shih, H. (1992). Beyond comprehension exercises in the ESL
academic reading class. TESOL Quarterly, 26(2), 289-
318.
Sjostrom, C. L., & Hare, V. L. (1984). Teaching high
school students to identify main ideas in expository
text. Journal of Educational Research, 78(2), 114-118.
Slater, W. H., Graves, M. F., & Piche, G. L. (1985).
Effects of structural organizers on ninth-grade
students' comprehension and recall of four patterns of
expository text. Reading Research Quarterly, 20(2),
189-202.
Slater, W. H., Graves, M. F., Scott, S. B., & Redd-Boyd, T.
H. (1988). Discourse structure and college freshmen's
recall and production of expository text. Research in
the Teaching of English, 22(1), 45-61.
Snow, M. A. (1993). Discipline-based foreign language
teaching: Implications from ESL/EFL. in H. Krueger &
F. Ryan (Eds.), Language and content: Discipline and
content-based approaches to language teaching (pp. 37-
56). Lexington, HA: D. C. Heath and Company.
Snow, H. A. (1994). Collaboration across disciplines in
postsecondary education: Attitudinal challenges. The
CATESOL Journal, 7(1), 59-64.
Snow, M. A., & Brinton, D. M. (1988). Content-based
language instruction: Investigating the effectiveness
of the adjunct model. TESOL Quarterly, 22(4), 553-574.
222
Snow, M. A., & Tricamo, J. (1991). Making access
meaningful: The language development general education
program for high risk language minority students.
Grant submitted to the Fund for the Improvement of
Postsecondary Education, U.S. Department of Education.
Los Angeles, CA: California State University, Los
Angeles.
Solarz, F. Perelman de. (1994). La construccidn del
resumen [The construction of the summary]. Lecture y
Vida, 15(1), 5-20.
Spivey, N, N. (1983). Discourse synthesis: Constructing
texts in reading and writing. Unpublished doctoral
dissertation. The University of Texas at Austin.
Spivey, N. N. (1987). Constructing constructivism:
Reading research in the United States. Poetics, 16,
169-192.
Spivey, N. N. (1990). Transforming texts: Constructive
processes in reading and writing. Written
Communication/ 7(2), 256-287.
Spivey, N. N., & King, J. R. (1989). Readers as writers
composing from sources. Reading Research Quarterly,
24(1), 7-26.
Strode, S. L. (1989). Re-evaluation of the effects of
annotation training on summary writing and
comprehension of college students. Unpublished
doctoral dissertation, University of Missouri, Kansas
City.
Swales, J. (1990). Genre analysis: English in academic
and research settings. Cambridge: Cambridge
University Press.
Taylor, B. M. (1982). Text structure and children's
comprehension and memory for expository material.
Journal of Educational Psychology, 74(3), 323-340.
Taylor, B. M., & Beach, R. W. (1984). The effects of text
structure on middle-grade students comprehension and
production of expository text. Reading Research
Quarterly, 19(2), 134-146.
Taylor, K. K. (1984). The different summary skills of
inexperienced and professional writers. Journal of
Reading, 27(8), 691-698.
223
Taylor, K. K. (1986). Summary writing by young children*
Reading Research Quarterly, 21(2), 132*149.
Terrell, T. D. (1991). The role of grammar instruction in
a communicative approach. The Modern Language Journal,
75, 52-63.
Tharp, R. 6., & Gallimore, R. (1988). Rousing minds to
life: Teaching, learning, and schooling in social
context. Cambridge: Cambridge University Press.
The Offices of Public Affairs and Analytical Studies.
(1994). Facts 13. California State University. Los
Angeles, CA.
The TOEFL test of written English guide. (1989).
Princeton, NJ: Educational Testing Service.
Tricamo, J., & Snow, H. A. (1994). Project LEAP: Learning
English-fOr-Academic-Purposes. Final Report to FIPSE.
Los Angeles, CA: California State University, Los
Angeles.
Urztta, C. (1987). You stopped too soon: Second language
children composing and revising. TESOL Quarterly,
21(2), 279-304.
Ushakova, T. N. (1994). Inner speech and second language
acquisition: An experimental-theoretical approach. In
J. P. Lantolf & G. Appel (Eds.), Vygotskian approaches
to second language research (pp. 135-156). Norwood,
NJ: Ablex Publishing Corporation.
Valdds, G. (1992). Bilingual minorities and language
issues in writing: Toward professionwide responses to
a new challenge. Written Communication, 9(1), 85-136.
Valentine, J. F., & Repath-Martos, L. (1992). How relevant
is relevance?: An examination of student needs,
interest, and motivation in the content-based
university classroom. The CATESOL Journal, 5(10), 25-
42.
van Dijk, T. A. (1980). Macrostructures: An
interdisciplinary study of global structures in
discourse, interaction, and cognition. Hillsdale, NJ:
Lawrence Erlbaum Associates, Inc.
van Dijk, T. A., & Kintsch, W. (1983). Strategies in
discourse comprehension. New York, NY: Academic
Press.
224
VanPatten, B. (1994). Evaluating the role of conscioueness
in eecond language acquisition: Terms, linguistic
features & research methodology. Applied Linguistics,
11, 27-36.
Waxman, H. C., Felix, Walker de, J., Martinez, A., Knight,
s. L., & Padrdn, Y. (1995). Effects of implementing
classroom instructional models on English language
learners' cognitive and affective outcomes. Bilingual
Research Journal, 18(3 & 4), 1-22.
Weinstein, C. E. (1982). Training students to use
elaboration learning strategies. Contemporary
Educational Psychology, 7, 301-311.
Weinstein, C. E. (1987). Fostering learning autonomy
through the use of learning strategies. Journal of
Reading, 30(7), 590-595.
Weinstein, C. E., & Mayer, R. E. (1986). The teaching of
learning strategies. In M. C. Wittrock (Ed.), Handbook
of Research on Teaching (3rd ed., pp. 315-327). New
York: Macmillan.
Wesche, M. B. (1993). Discipline-based approaches to
language study: Research issues and outcomes. In M.
Krueger & F. Ryan (Eds.), Language and content:
Discipline and content-based approaches to language
teaching (pp. 57-79). Lexington, MA: D. C. Heath and
Company.
Western Association of Schools and Colleges. (1990).
Accreditation report. Los Angeles, CA: California
State University, Los Angeles.
Wilhelm, R. (1993). Summarizing text for intelligent
communication. Paper presented at the International
Conference and Research Center for Computer Science,
Schloss Dagstuhl.
Winograd, P. (1984). Strategic Difficulties in summarizing
tasks. Reading Research Quarterly, 19(4), 404-425.
Witte, S., & Faigley, L. (1981). Coherence, cohesion, and
writing quality. College Composition and
Communication, 32(2), 189-204.
Zamel, V. (1983). The composing processes of advanced ESL
students: Six case studies. TESOL Quarterly, 17(2),
165-187.
22S
Zimmerman, C. B. (1994)* Self-selected reading and
Interactive vocabulary Instruction: Knowledge of
perceptions of word learning among L2 learners.
Unpublished doctoral dissertation, University of
Southern California.
APPENDIX A
Student Profile Questionnaire
The purpose of this questionnaire is to collect information about students enrolled in
your dass. Your identity will be protected and will not be revealed. All individual
responses will remain CONFIDENTIAL and only group results will be reported.
Please answer ALL of the questions.
INSTRUCTIONS: Fill In the spaces below or mark the boxes with an "X” where
appropriate.
Background Information
(1) In what country were you born?
□U .S. □Korea □South America.
□ Mexico □Another country in specify_______
□Central America, Asia, specify_______ □Vietnam
specify___________ □Pacific Islands, □Other, specify________
□China specify_______
(2) If you were not born in the U.S., how long have you lived in the U.S.?
□Less than 2 years □Between 8 and 10 years
□ Between 2 and 4 years □More than 11 years
□ Between 5 and 7 years □ Since birth
(3) To what ethnic group do you belong?
□Asian □Mexican
□ Aslan-American □ Mexican-American.Chicanofa
□Caucasian □Pacific-lslander
□Latino/a □Other, specify_____________
(4) Age on your last birthday
□ 20 or below □Between 21 and 25 □Above 25
(5) Gender............................................................................................. QMale □Female
(6) Class in college ..........................□Freshmen □Sophomore □Junior □Senior
(7) Are you an EOP (Educational Opportunity Program) student? □Y es QNo
(8) Are you an international student? To be an international student,
you need to be studying here on a student visa .......................................QYes QNo
Have you taken any of the following classes at CSLA: Check ALL that apply.
(9) EDCI0937 .................................................................................................QYes QNo
(10) English 0 9 5 ? .............................................................................................. OYes QNo
(11) English 0 9 6 ? .............................................................................................. QYes QNo
(12) English 101 or English 190?..................................................................... QYes QNo
227
Are you currently enrolled in any of the following classes at CSLA:
(13) EDO 0937 ....................................................................................................QYes C)No
(14) English 095 ....................................................................................................QYes QNo
(15) English 0967 ................................................................................................. QYes QNo
(16) English 1 0 1 7 ................................................................................................. QYes QNo
Language and School Background
(17) What was the FIRST language you spoke when you were a child in your home?
□ English O Vietnamese □ Other,
□Spanish □ Filipino language specify_
□Cantonese □ Mandarin □ Does not apply
□Korean
If your answer to question 17 w as ENGLISH, STOP HERE. If your answer
w as a language OTHER THAN ENGUSH, COMPLETE THE REMAINDER OF
THE QUESTIONNAIRE.
(18) What OTHER language/s did you speak when you were a child-BEFORE you started
school? Check ALL that apply.
□ English □ Vietnamese □ Other,
□Spanish □ Filipino language specify___________
□Cantonese QMandarin □ Does not apply
□Korean
(19) What language do you most frequently use NOW?
□ English □ Vietnamese □ Other,
□ Spanish □ Filipino language specify___________
□Cantonese QMandarin □ Does not apply
□Korean
(20) What schooling have you received in the U.S.? Check ALL that apply.
□Kindergarten through college QHIgh school through college
□Elementary school through college □ Community college
□Junior high school through college QNone of the above
(21) What type of language programs have you attended in the U.S.? Check ALL that
apply.
□ESL (English as a second language! QOnly English classes
classes □ Other, specify_______
□ Bilingual classes
How well would you say you are able to do the following in ENGLISH:
(22) Speak English?......................................................... QVery well QWell QNot well
(23) Understand spoken English?.................................. QVery well QWell QNot well
228
(24) Read E nglish?.......................................................... □Very well QWell QNot well
(25) Write English?......................................................... QVery well QWell QNot well
How well would you say you are able to do the following in ENGLISH at CSLA:
(26) Speak English in c la s s ............................................ QVery well QWell QNot well
(27) Understand class lectures in English..................... QVery well QWell QNot well
(28) Read school textbooks in English.......................... QVery well QWell QNot well
(29) Write for academic purposes in English QVery well QWell QNot well
How well would you say you are able to do the following in the NON-ENGLISH language:
(30) Speak the non-English language?.......................... QVery well QWell QNot well
(31) Understand the non-English language? ................ QVery well QWell QNot well
(32) Read the non-English language?............................. QVery well QWell QNot well
(33) Write the non-English language?.......................... QVery well QWell QNot well
Language Use
In what language do you do the following? Check ALL that apply.
English Language
Other
than
English
Does
not
apply
(34) Talk to your parents/uncles/aunts?...................... Q Q Q
(35) Talk to your husband/wife/boy/girl-friend? . . .
Q Q Q
(36) Talk to siblings?..................................................... Q Q □
(37) Talk to children? ..................................................
Q a Q
(38) Talk to friends socially?........................................
Q Q Q
(39) Talk to classmates?................................................
Q Q Q
(40) Buy luxury g o o d s.................................................. Q Q Q
(41) Do the grocery shopping?.....................................
Q Q Q
(42) Listen to the radio?................................................
Q Q Q
(43) Watch T V ? ............................................................
Q Q Q
(44) Read newspapers?..................................................
Q Q Q
(45) Read magazines? ..................................................
Q Q
Q
(46) Read books? ..........................................................
Q Q Q
(47) Attend religious services?.....................................
Q Q Q
229
English Language
Other
than
English
(48) Take tests (such as DMV driving test)?............. □ □
(49) Talk to your doctor?............................................ □ □
Does
not
apply
□
□
230
APPENDIX B
EQUIVALENCE IN CONTENT KNOWLEDGE
QUESTION FOR HEALTH SCIENCE ISO
LAST NAME FIRST NAME M. I.
In FIVE minutes, write all that you know about GENDER DIFFERENCES IN
HEART DISEASE. Use words, phrases, or sentences. DO NOT WORRY ABOUT
USING PERFECT GRAMMAR.
231
QUESTION FOR CULTURAL ANTHROPOLOGY 250
LAST NAME FIRST NAME M. I.
In FIVE minutes, write ell that you know about CULTURAL DIFFERENCES IN
MURDER HABITS. Use words, phrases, or sentences. DO NOT HORRY ABOUT
USING PERFECT GRAMMAR.
232
QUESTION FOR ANIMAL BIOLOGY 155
LAST NAME FIRST NAME M. I.
In FIVE minutes* write all that you know about THE SUN AND ITS EFFECT ON
THE BODY RHYTHMS. Use words, phrases, or sentences. DO NOT WORRY ABOUT
USING PERFECT GRAMMAR.
233
APPENDIX C
HANDOUT! Buaaar; Writing, Awamaii of Task Demands
INSTRUCTIONSi Read the article and the model summary. Using the
Evaluation Checklist for Summary Writing below analyse the model
summary.
The author!
1. Has deleted minor details. YES NO
Give examples.
2. Has combined ideas. YES NO
Give examples.
3. Has paraphrased accurately. YES NO
Give examples.
4. Has reflected the author*■ emphasis. YES NO
Give examples.
S. Has identified the topic. YES NO
Give examples.
6. Has identified the main idea. YES NO
Give examples.
7. Has stayed within appropriate length. YES NO
Give examples.
8. Has excluded personal opinion. YES NO
Give examples
9. Has shown that the summary is based
on other people's material. YES NO
Give examples.
Adapted from Casazza, H. E. (1993); Kirkland, H. R., 6 Saunders, H. A.
(1991).
234
APPENDIX D
HANDOUTI STRATEGIES POR SUMMARY WRITING
SUMMARISING PROCESS
Aftar raading tba article to bo aunariiadi
1. Put tho orticla down and thinkt
What ia/aro tha wain idaa/a praaantad in tha
article?
A* Maka a mental picture of tho article's main idea/s.
B. Write down tha main idea/s IN YOUR OWN WORDS.
2. Delate redundant and minor details.
A. Do NOT include the same information more than once.
B. Do NOT include unimportant details.
3. Combine ideas in different paragraphs or sections of
the article.
A. Do NOT write one sentence for every paragraph.
B. Combine IMPORTANT information from different
paragraphs or sections of the article.
4. Combine lists.
A list of activities like "ice skating and skiing" can
be summarised by writing/saying "winter sports."
5. Go over your summary whan you have finished it.
A. Read your summary, read the article.
B. Is your summary complete? Does it emphasise the
points that the author emphasised in his/her
article?
6. Polish your summary.
A. Is your summary easy to follow?
The sentences/paragraphs in your summary should be
connected with words like "furthermore,"
"moreover," "in addition."
YOUR SUMMARY SHOULD
1. Show that it is based on other people's material.
Mention the author of the article periodically.
2. Avoid plagiarism. To avoid plagiarism you shouldt
A. Paraphrase (USB YOUR OWN WORDS) what the author has
stated OR
B. Use quotation narks if you use the author's words.
Then, write the author's last name and year of
publication in parenthesis [e.g., (Smith, 1991)].
3. Present "the big picture." Do NOT worry about details.
Adapted from Brown, A. L., & Day, J. D. (1983); Brown, A. L., Day, J.
D., St Jones, R. S. (1983); Hare, V. C., & Borchardt, K. M. (1984);
Kirkland, M. R., 4 Saunders, H. A. (1991).
235
APPENDIX E
HANDOUT! BIGHT STEPS TO EVALUATING A SUMMARY
Tha author
1. Identified tha main ldaa/a. YES
(Pay attention to tha titla and thinks
What wan tha author'a message?)
Suggaationa for improvement?
2. Delated minor dataila. YES
Suggaationa for improvement?
3. Reflected tha author'a emphaaia. YES
Did tha author focua on tha main idaaa?
(Tha titla may help identify tha
emphasis.)
Suggaationa for improvement?
4. Combined aimilar idaaa. YES
Suggaationa for improvement?
5. Paraphraaad accurately. YES
Remember that quotation* mark* had to be uaad
or the author'a idea* had to be paraphraaad.
Suggaationa for improvement?
6. Excluded peraonal opinion. YES
Tho writer ehould have avoided uaing
"X- "We."
Suggestions for improvement?
7. Showed that the summary ia based on YES
other people's material. (Tha author
uaad phraees auch aa "According to,"
"As atated by...")
Suggestions for improvement?
8. Connected the sentences with YES
words like Moreover, similarly,
in contrast, first, second.
Suggestions for improvement?
Adapted from Brown, A. L., & Day, J. D. (1983); Brown, A. L
D., 8 Jones, R. S. (1983); Hare, V. C., & Borchardt, K. H.
Xirkland, M. R., f i Saunders, M. A. (1991).
NO
NO
NO
NO
NO
NO
NO
NO
., Day, J.
(1984);
236
APPENDIX F
HANDOUT! CONNSCTXNO WORDS
A. To introduce additional idaaai
Also, Furthermore, Moreover, In addition,...
Add your ownt
B. To introduca a coapariaon (similarity)t
Similarly, Also, Both...and, Not only....but also,
Xiikewise,.•.
Add your ownt
C. To introduca contrast (diffarancss)t
On tha othar hand, Bowavar, In contrast, Yat, Whila,
Xnstaad, On tha contrary,...
Add your ownt
D. To introduca a summary or conclusion!
Zn conclusion, Xn summary. To conclude,
To summarise,...
Add your ownt
B. To show sequence of events or chronological ordert
First, Sscond, (Third, etc.; Next, Last, Finally...
Add your ownt
Oshima, A., & Hogue, A. (1991). Writing academic English.
Menlo Park, CA: Addison Wesley.
237
APPENDIX 0
THE SAFER SEX? HEALTH SCIENCE 150
Probing a cardiac gender gap
"Fifteen yeara ago, my mother died quite prematurely of heart
disease. Looking back, X have to feel that there was failure, from
every corner to help her," eaid Constance Horner, under secretary at the
Department of Health and Human Services, addressing a 1989 conference on
women and heart disease.
Horner's story offers a poignant reminder of the traditional and
often dangerous tendency to consider heart disease an affliction of men.
That assumption was especially prevalent several decades ago, when
physicians largely ignored women who complained of chest pain (angina),
a potential sign of impending heart attach.
Host people know that heart disease represents the number-one
killer of men in the United States. But many don't realise that it's
also the number-one killer of women. Although the hormone estrogen
apparently helps many women ward off coronary artery disease during
their reproductive years, this protective edge begins to vanish with
menopause. From then on, a woman's risk of developing heart disease
mounts rapidly, soon equaling that of men. Statistics clearly bear this
out* Between age 45 and 65, approximately one in nine U.S. women shows
symptoms of heart disease, compared with about one in seven men; after
age 65, that ratio humps to one in three, roughly matching the ratio in
men.
But while the disease gap eventually disappears, another gap lurks
in the statistics! a disquieting difference between men's and women's
cardiac aoetmlity rates. During the first few weeks after a heart
attack, women face twice the risk of death men, and their heightened
relative risk persists well into the following year. In addition,
female candidates for coronary artery bypass surgery must confront the
stark fact that the operation itself— a risky but potentially lifesaving
procedure— puts women in greater jeopardy of death than men.
Does this mortality gap reflect some innate, biological difference
in the female body? Are the statistical discrepancies due to known risk
factors, such as women's tendency to suffer heart attacks at a later
(and presumably frailer) age than men? To what extent do women or their
physicians overlook early symptoms of heart disease, delaying important
treatment?
Recently published research and findings presented in November at
the American Heart Association (AHA) scientific sessions in Dallas
provide conflicting data and few answers. Nevertheless, many
investigators view these collectively murky results as a first step
toward understanding the vagaries of heart disease in women.
In the 1980s, several research reports attributed women's higher
death rate from heart attacks to the fact that women typically suffer
attacks about a decade later in life than men, and thus may not recover
as readily as a younger person of either sex. But new data from the
University of Massachusetts Medical School in Worcester indicate that
the deadly difference doesn't spring solely from age.
Cardiologist Richard C. Becker and his colleagues studied 2,742
men and 597 women who were hospitalised for a first or second heart
attack, caused by a clot that restricted blood flow to the heart. All
patients received a clot-busting drug (recombinant tissue plasminogen
activator) within four hours after the attack. At the AHA meeting,
Becker reported that 9 percent of the women and 4 percent of the men
died during the initial hospital stay. And among those who survived the
attack, 12 percent of the women died during the following year
238
(primarily from cardiovascular disease), compared with 6 percent of the
men.
The researchers found that the women in the study were more than
twice as likely as the men to develop chronic heart failure after the
attack. In people with this potentially lethal condition, lingering
damage from the acute attack permanently impairs the heart's pumping
ability. During the first few weeks after the heart attacks, 5.7
percent of the women and 2.1 percent of the men developed chronic heart
failure, Becker ^reports.
The patients* ages in this study followed a familiar pattern! 24
percent of the women, and only 9 percent of the men, were more than 70
years old. And the age distribution did make a difference in the
mortality risk, Becker says- but it didn't account for the entire gender
gap.
In the statistical assessment of risk factors, the researchers
also isolated preexisting, chronic conditions that could complicate a
heart patient's recovery. Compared with the men in the study, the women
had a higher prevalence of high blood pressure and diabetes, they found.
But even when the team statistically controlled for these conditions,
lumping them in with the age factor, a significant mortality gap
remained.
Becker says he can't explain the risk remnant, but he and others
speculate that differences in female physiology might somehow tip the
scales. "The heart of a woman doesn't tolerate a myocardial infarction
(heart attack] as well as that of a man," ha suggests.
Another report, however, argues against the notion of a built-in
risk for women. Nicholas H. Fiebach and his colleagues at Yale
University studied 332 female and 790 male heart attack sufferers,
findings that 14 percent of the women and 9 percent of the men died
during the initial hospitalisation. When the researchers controlled not
only for the risks associated with age, diabetes and hypertension, but
also for the severity of heart disease, they managed to account for all
but 1,5 percent of the gender gap--an insignificant remnant, Fiebach
says.
The Yale researchers conclude that the severity factor merits
special attention in efforts to explain women's heightened mortality
risk.
Several of their findings suggest that women tend to reach the
cardiac care unit with more serious heart conditions than men, they
report in the Feb. 23, 1990 JOURNAL OF THB AMERICAN MEDICAL ASSOCIATION.
For example, hospital admitting exams showed that the women in this
sample, compared with the men, had more severely impaired left
ventricular function--an important measure of heart attack severity.
Fiebach believes some physicians may overlook cardiac warning
signs in women, especially if those patients are relatively young. "If
a middle-aged woman is having chest pain, one doesn't think of a heart
attack. If a middle-aged man is having chest pain, one automatically
thinks of a heart attack," he told SCIENCE NEWS. He also suggests that
heart disease may be more difficult to diagnose in women than in men.
In any case, a failure to diagnose and treat coronary artery disease
until it becomes more severe could lead to more damaging heart attacks,
which in turn carry a higher risk of death, he says.
Cardiovascular researcher, Kathleen B. King agrees that the
traditional diagnostic bias still influences medical thinking. "There's
definitely an attitude that heart disease is a man's problem," says King
at ths University of Rochester in New York.
And physicians may not be alone in their delayed response to early
warning signs, or even an outright heart attack, in women. Fiebach
found that the women in his study generally waited longer than the men
to head for the hospital after their chest pains began. "It may be that
239
tha longer you wait, the more chances are that you*re going to have a
sever complication," he suggests, while cautioning that the link between
such delays and a poor clinical outcome remains speculative.
The Tale report also hints that physicians may treat male and
female heart attack patients differently once their stay in the hospital
has ended. The researchers noted that physiciane prescribed beta-
blocker drugs— which alow the heartbeat and may improve survival outlook
for some peopla— more often for men than for women at the time of
hospital discharge. Fiebach emphasizes, however, that this finding did
not reach statistical significance in the study. Further research is
needed, he says, to determine whether such gender-based treatment
differences exist.
The heart attack itself is only part of the story. Researchers
also see disturbing differences between the sexes when they look at
mortality rates during and after coronary artery bypass surgery, in
which surgeons attempt to restore blood flow to the heart by attaching a
vein above and below the site of a coronary artery blockage to shunt
blood around it. Whether undertaken before or after a heart attack,
bypass surgery can somatimas save lives— but statistics show that the
operation entails a greater gamble for women than for men.
A research team led by cardiologist Steven S. Khan at the
University of California, Los Angeles, has uncovered evidence suggesting
that woman with heart disease tend to get bypass surgery later in the
course of their disease. Khan attributes such delays, which may
increase the riskiness of the surgery, to a "referral bias" on tha part
of physicians.
"Women pretty much have to have a heart attack or very severe
symptoms— life-threatening symptoms of their heart disease— before they
are sent to surgery," Khan says. Like other researchers, he suspects
that some women who complain of chest pain get the brush-off by
physicians who think the pain is psychosomatic. He also notes that
heart disease may be trickier to diagnose in women, whose reports of
chest pain don't always match the traditional description of angina.
Moreover, physicians may worry about advising some women— especially
those debilitated by old age or suffering serious complicating condition
— to undergo the rigorous bypass procedure unless thair heart condition
seems dire.
Khan and his colleagues assessed the mortality rates of 1,815 men
and 482 women who had bypass surgery between 1982 and 1987. They report
in the April 15, 1990 ANNALS OF INTERNALS MEDICINE that 4.6 percent of
tha women, compared with 2.6 percent of the men, died during surgery or
the postoperative recovery period. But when the researchers
statistically controlled for age and severity of illness, the mortality
gap vanished.
In general, they found, the women ware sicker than the men when
wheeled into the operating room, showing more severe signs of heart
disease, including a particularly risky type of chest pain called
unstable angina. Regardless of gender, says Xhan, a patient's odds of
dying during or after bypass surgery increase with the severity of his
or her heart condition.
The notion that physicians refer only the sickest women for
surgery finds no support in a study conducted at Downstate Medical
Canter in Brooklyn, N.I. Thierry A. Folliguet and his co-workers
assessed several measures of heart disease severity— including chest
pain and the number of blocked arteries— in 858 men and 439 women who
underwent bypass surgery at three New York hospitals in 1989. At the
AHA meeting, Folliguet reported a total lack of evidence indicating that
the women were more severely ill than the men when they reached the
operating room. He says he cannot explain the discrepancy between his
finding and Khan's.
240
Folliguet favors a long-standing explanation of women's higher
mortality rate from bypass surgeryt body Bias. Because women typically
are smaller than men, he says, their bypass operations pose more
technical difficulties. Many cardiovascular surgeons espouse this view,
asserting that smaller arteries reduce a woman's chance of surviving the
surgery and recovery period.
Other researchers have uncovered data contradicting the body-sixe
theory. King, for instance, focused on 464 women who had bypass surgery
between 1983 and 1988, finding that 4.3 percent of them died during or
soon after the operation. At tha AHA meeting, she reported that a
statistical analysis comparing the women who died with those who
survived showed no link between bypass risk and body surface area.
King quickly points out that body surface area, calculated form
height and weight, may not correlate with tha sixe of coronary arteries.
She and her colleagues want to conduct a follow-up study measuring blood
vessel else and tracking mortality rates, in the those of establishing
whether or not small vessel sixe puts women— and smaller men— at extra
risk during bypass surgery.
The Rochester results also punch holes in a theory attributing
women's higher bypass risk to their tendency to suffer heart attacks at
an older age than men. King's all-female sample showed no correlation
between age and bypass mortality. "Our study showed that age really
didn't make a difference," she told SCIENCE NEWS.
Potential complicating factors, such as diabetes or high blood
pressure, similarly gained to explain these woman's risk, aha says.
Many heart researchers today suspect that a complex interplay of
factors— including age, severity of illness and perhaps some innate
biological differences— works against women who suffer heart attacks.
But that risk equation remains largely unresolved.
"We're just beginning to explain how all these things contribute,"
Fiebach says.
While no amount of research can save her mother now, Horner points
out that soma 247,000 U.S. women face the prospect of a deadly heart
attack in tha coming year. Given the magnitude and urgency of that
threat, she urges researchers to pursue their investigations with vigor.
In the meantime, one fact is clean Men have no monopoly on
haxards of the heart.
Fackelmann, K. A. (1991). The safer sex: Probing a
cardiac gender gap. Science News, 139(3), 139-140.
241
MURDER IN GOOD COMPANYt CULTURAL ANTHROPOLOGY 250
Cooperation, camaraderie and a dizzying homicide rate distinguish a
small New Guinea society
The Gebusi, a society of about 450 persons living in a New Guinea
rain forest, are a strikingly gentle lot. They revel in Aog-wa-yay,
roughly translated as "good company." Togetherness, casual talk and
exuberant humor are daily staples. There is no central political
structure and no jockeying for power among the stronger men; matters of
concern to Gebusi, who live in communal "longhouse" settlements, are
decided by consensus. Pood, including bananas grown in small gardens
and the occasionally hunted wild pig, is routinely shared among all the
residents of a settlement. Anger, violence and warfare are frowned
upon.
But behind this aura of serenity and conviviality lurks a brutal
paradox; The Gebusi murder ona another at a rate among the highest ever
reported, about 40 times greater than the 1980 homicide rata in tha
United States. According to anthropologist Bruce M. Knauft of Emory
University in Atlanta, who documented the Gebusi homicide rate during
nearly two years of field work, their murderous ways cannot be explained
by current theories of violence applied to societies such as ours that
have complex political and economic systems.
"The character of homicide appears to change in simple societies
that have no pecking order or dominance hierarchy among adult men," says
Knauft. "Especially in these societies, there may be a patterns of
social life that is generally peaceful and tranquil but is punctuated by
aggression which, when it does occur, is unrestrained and frequently
homicidal."
This pattern, he adds, may be a critical aspect of the evolution
of human violence, since simple, decentralised societies have
predominated for most of the history of Homo sapiens.
Anthropologists who encounter such societies often come away with
a vision of "Eden in the outback,” says Emory anthropologist Melvin
Konner, "only to have the discovery foiled by better data." Margaret
Head's observations of life in Samoa are perhaps tha moat famous
instance of this tendency. In 1928, she wrote of a near-utopian Samoa
unencumbered by aggression, competition, sexual repression and conflict-
ridden adolescence. Since the 1950s, several accounts, including a
controversial book by Australian anthropologist Darek Freeman, have
described a more complex Samoa in which violent crime, including rape,
is not uncommon and is often committed by adolescent boys.
In small societies of noncompetitive foragers or hunter-gatherers,
violence is especially easy to overlook, says Knauft. First, the people
often fear and downplay occasional aggressive outbursts, and second,
even a few murders committed from year to year can translate into a high
overall homicide rate.
A case in point are the IKung Bushmen of the Kalahari Desert,
dubbed "the harmless people" by an investigator several decades ago.
They are indeed generally peaceful and gregarious, but a systematic
survey in 1979 found that the IKung's homicide rate is nearly three
times that of the United States, which is already one of the highest
among Hestarn nations. Within a population of 1,500 IKung, an estimated
22 killings occurred over five decades, "about five more than the name
number of New Yorkers would have been expected to commit over the same
period," according to Konner.
A similar pattern of pervasive good will and self-effacement
combined with occasional violent flare-ups and even murders has been
noted among Central Eskimo groups, the Semai aborigines of Malaysia and
tho nomadic Hadsa of Tanzania, says Knauft.
242
But the Gebusi study is perhaps the most intensive homicide
inquiry to date. Knauft beqan by establishing complete genealogies for
IS of 2S Gebusi clans and partial genealogies for three others. Clan
members are distantly or directly related to one another, the total
number of adult deaths in the genealogical sample between 1940 and 1982
totaled 394* The cause of each death was cross-checked in extensive
discussions with Gebusi informants, including relatives, friends and
acquaintances of the deceased person.
There were 129 cases of homicide, nearly one-third of all deaths,
reports Xnauft in the Aug-Oct. CURRENT ANTHROPOLOGY. The annual
homicide rate from 1940 to 1982 is at least 868 murders per 100,000
persons, he says, a conservative calculation based on the partial
population survey. By contrast, the 1980 homicide rate in the United
States, according to Federal Bureau of Investigation statistics, was
10.7 murders per 100,000; the estimate for Detroit in 1988 was 88.2
murders per 100,000 persons.
"Only the more extreme instances of modern mass slaughter would
equal or surpass the Gebusi homicide rate over a period of several
decades," says Knauft.
Four out of five Gebusi murders uncovered in the study involved
the killing of someone branded as a sorcerer for having allegedly caused
the death of another Gebusi. Death from diseases often caused by
infections and parasites is ever-present among the Gebusi and sets tho
stage for sorcery accusations. Hore than one-quarter of the sickness
deaths in the genealogical sample precipitated a sorcery killing.
Although sorcery accusations are leveled at both man and women,
only man carry out the homicides, says Knauft "Including some who were
among the least assertive and aggressive even by Gebusi standards." In
two communities he surveyed, 11 of 17 adult men had killed an alleged
sorcerer.
The Gebusi believe that a sorcerer conceals his or her evil
intentions toward others. After a sickness death, the first step in
uncovering a possible sorcerer is to hold an all-night seance conducted
by mediums who believe their bodies are temporarily inhabited by
spirits. If the seance provides a sorcery suspect, that person must
pass a "divination inquest." In many cases, the accused must cook large
piece of fish or meat coated with powder ground from sago palms without
burning it. Fires are placed around the food, rather than under it,
making proper cooking more difficult. Cooking is performed on the
hearth of the dead person, whose spirit is thought to place its hands
underneath tha food to ensure proper heating for innocent sorcery
suspects.
If the outcome is damning, the suspect may be killed on the spot,
with no intervention by his or her kin. In some cases, says Knauft,
close kin are ambivalent about a suspect's guilt, causing accusers to
wait waeks or months before they organise a group to ambush the alleged
sorcerer in the forest away from supporters.
Revenge killings, even after an ambush, are rare, notes Knauft.
Only four of the reported sorcery murders resulted in a return killing
by relatives of the slain person. Host killings in recent years, he
points out, have taken place with the implied consent, mainly through
lack of vocal support, of all the accused sorcerer's close kin.
Otherwise, there is the possibility that a grieving relative could
report the murder to New Guinea government authorities. A couple of
cases, both involving the killing of an accused sorcerer in a fit of
rags, have resulted in reports to state officials and Imprisonment of
the killers.
A few other killings in Knauft*s sample occurred during
ritualistic bow-and-arrow skirmishes between Gebusi from different
settlements.
243
Add it all up, and tha peaceful Gebusi have quite a bit of blood
on their hande. But eince sorcery is probably the last thing any Gebusi
of sound mind would practice, why are so many murdered for their alleged
witchcraft?
Current theories of violence offer no satisfactory answers,
contends Knauft. For example, sociobiologists have proposed that one
characteristic of violence in human societies is a kind of "genetic
selection.” Zn this view, which draws on the notion that there is an
innate tendency toward aggression, an offender is more likely to murder
a genetically unrelated victim In order to minimise the loss to the gene
pool of persons who share a high percentage of his or her genes.
Gebusi violence, however, seems to defy this prediction. Zn a
breakdown of genetic relationships in one community of 30 men and 34
women, Knauft found that homicide victims were far more likely to be
killed by relatives living in the same longhouse settlement.
Another theory, springing from the position that violence is
learned, holds that local groups of related males in simple societies
compete and fight with other "fraternal interest-groups." These hostile
bands are said to be precursors of military organisations in larger,
more complex societies.
But aggressive, man-only organisations are lacking in Gebusi
society, explains Knauft. Zn fact, married men often move to the
settlements of their in-laws, creating overlapping social ties
throughout the society. Zn general, men as well as women go out of
their way to share food and other resources. Displaying dominance or
superiority over others is treated as a breach of "good company."
A third possibility, based on several cross-cultural studies, is
that an emphasis on punishmsnt and obedience in child-rearing by
authoritarian, emotionally distant fathers promotes aggressive adult
male behavior. Here again, says Knauft, tha Gebusi are theory busters.
Physical punishment of children in their society is extremely rare.
Fathers are affectionate and undemanding of their sons, he observes,
"and in two years Z only once saw a father so much as shove his son in
irritation.” The greatest threat to a child's trust is the stark
realisation that mother or father may soon die, either from natural
causes or at the hands of sorcery accusers.
According to Knauft, in many cases "sorcery homicide is ultimately
about male control of marriageable women." Although Gebusi rarely
discuss the balance of women in a clan or smaller living groups, there
is an underlying expectation that after a woman marries, a close female
relative of her husband's will at some point be encouraged to marry back
into the woman's clan of origin. A strikingly high number of sorcery
accusations and murders, says Knauft, are aimed at members of a group
thought to be holding out on its obligation to promote marital exchange.
Zt is not unusual, he notes, for a woman to be accused of sorcery and
killed by her own brothers as a way of taking vengeance on their
brothers-in-law.
Nevertheless, the Gebusi deny that there are male disputes over
women. "The heavy cultural pressure against anger expression seems to
maximise the problems in their system of marital exchange," says Knauft.
Zn simple societies such as the Gebusi, marked by a lack of male
status distinctions, Knauft proposes that violence revolves around the
control of sexual relations and that sporadic, violent incidents serve
to reestablish cooperation in the community.
A number of anthropologists welcome Knauft's extensive homicide
data but add their own caveats or interpretations to the findings.
Both Leland Donald of the University of Victoria, British
Columbia, and Robert K. Dentan of the State University of New York in
Buffalo suggest that a history of being raided by more powerful
neighboring societies contributes to an attitude of helplessness, rather
244
than harmleesness, among people like tho IKung, Semai and Gebusi.
Resignation to the inevitability of periodic massacres, they say, can
encourage a cultural retreat from angry and violent expressions.
Semai communities are also being pressured by the Malay government
to give up their land for industrial development and move to urban
centers, says Dentan, who has done field work among the Semai. Although
tensions are building and Semai murders, "either of arrogant outsiders
or of friends in drunken outbursts,” are likely to increase, Dentan has
no evidence of periodic Semai homicide related to male rivalry over
marriageable women.
"Knauft*s account is probably tha best description of capital
punishment in the ethnographic literature,” comments Keith P. Otterbein
of the State University of New York in Buffalo, Gebusi killings, he
says, are usually viewed as legitimate, seldom involve retaliation and
involve the premeditated execution of a suspect found guilty, elements
that define capital punishment. Otterbein notes that sudden, unplanned
killings occur more often among tha Eskimo and IKung, a homicidal
variation that Knauft does not account for.
Knauft, however, says tha line between accepted and inappropriate
killings in simple societies "can become vanishingly thin." In the case
of the Gebusi, some sorcery murders are accepted by the community while
others are actively opposed or tolerated.only our of cowardice or fear
by the victim's supporters.
There is also a vanishingly thin line between Knauft*s explanation
of Gebusi homicide and sociobiological theory, write Martin Daly and
Margo Wilson of McMaster University in Hamilton, Ontario, in a
commentary accompanying Knauft'a article. Pegging sorcery killings to
male control of marriageable women, they contend, echoes the
sociobiological emphasis on male competition to reproduce with available
females as an important spur to violence and homicide.
According to Daly and Wilson, "it is clear that homicides, among
the Gebusi as among other peoples, arise out of conflicts of interest."
They also question whether Knauft's reliance on the recollections
of Gebusi individuals, without having observed for himself any
homicides, has distorted his findings.
Extensive cross-checking of each homicide account, as well as the
details of associated seances and divinations, produced a relatively
accurate, conservative estimate of tha murder rate, responds Knauft. On
two occasions, he adds, his presence and that of his colleague and wife
Eileen M. Cantrell may have prevented a sorcery killing. In one case,
the anthropologists gave verbal support to the relatives of a woman who
was being threatened by her accusers. It was a classic ethical
challenge for those studying homicide. "We couldn't just stand by and
let someone be killed,” says Knauft.
Bower, B. (1988). Murder in good company. Science News,
233(5), 90-91.
245
HERB COMES THE SUN! ANIMAL BIOLOGY 1SS
Scientists shod now light on winter depression
A young woman lies asleep on an overcast winter morning. At 4
a.m., a faint low emanates from a light bulb placed near her bed. The
glow gradually gains intensity and bathes the room in soft light by 6
a.m., when the woman awakens.
She has just welcomed the simulated dawn of a new day. After
several mornings brightened with light designed to mimic a bona fide
sunrise, the clouds begin to lift from the woman's "winter depression",
which appears punctually in late November of every year and abates by
the following April.
Preliminary support for the beneficial "sleeper effect” of
simulated dawns adds an intriguing new twist to the much investigated
thaw of winter depression following one or two weeks of daily exposure
to bright lights, in fact, as the study of seasonally recurring
depression— known by the apt acronym SAD, for seasonal affective
disorder— enters its second decade, researchers possess an abundance of
illuminating new findings that, in true scientific fashion, raise
complex new questions.
Kuch of the latest research was presented at the recent annual
meetings of the Society for Light Treatment and Biological Rhythms in
Bethesda, Kd., and the American Psychiatric Association in Washington,
DC.
Along with feelings of sadness, anxiety, and lethargy, people
experiencing SAD often display symptoms not typically seen in cases of
nonseasonal depression. These include difficulty awakening in the
morning, daytime drowsiness, cravings for sweat or starchy carbohydrate
foods, and extreme weight gain. Social withdrawal and a marked drop in
work performance also characterise SAD, which usually occurs for the
first time in people in their early 20s.
A large segment of the population suffers from winter depression,
with the numbers increasing in regions farther form the equator, where
winter nights grow longest (SNi9/23/89, p. 198). Several independent
studies estimate that the condition affects 1 to 2 percent of Florida's
population, about 6 percent of people living in Maryland and New York
City, and nearly 10 percent of the residents of New Hampshire and
Alaska. The prevalence of mild SAD symptoms ranges from nearly 3
percent in Florida to 11 percent in New Hampshire.
About IS percent of those hospitalised for severe depression
display a regular worsening or reappearance of symptoms in the winter,
according to studies conducted in the United States and Europe.
However, many SAD sufferers never seek mental health care, apparently
deciding to slog through each winter in anticipation of a corrective
dose of spring sunshine.
But numerous studies conducted over the past decade promote a
faster SAD-busting techniquei daily exposure to bright lights for a
week or two during the winter. Researchers currently favor two
approaches to this treatmenti either placing an individual in front of a
screen emitting light five times brighter than ordinary room light for
two hours each day, or using a screen producing light 20 times brighter
than normal room light for 30 minutes each day. Horning sessions often
yield the best results, but evening exposures can also provide relief.
Although the proportion of SAD patients whose symptoms disappear
in response to bright-light therapy varies widely from study to study,
the majority of reports cite strong improvement in at least 60 percent
of participants, assert Michael Terman and Jiuan Su Terman,
psychologists at the New York State Psychiatric Institute in New York
city, in a January 1991 review prepared for the federal government's
246
"depression guidelines panel.” A much smaller proportion of seasonally
depressed people feel markedly better after placebo treatment with dim
lights, the researchers add.
Observations of changes in the daily biological clock# or
circadian rhythms# of rats exposed to bright light at specific times
inspired the first attempts to treat winter depression with lights.
Scientists have no uncontested theory to explain how bright lights help
human SAD sufferers# but most assume that the treatment must somehow
alter circadian rhythms. One theory holds that SAD often derives# at
least in part# from the delayed nightly secretion of melanin# a hormone
involved in the regulation of sleep; some evidence suggests that
successful bright-light therapy can initiate melatonin secretion an hour
or two earlier. Scientists refer to this readjustment of the biological
clock as a "phase advance.”
Simulated dawns may also prompt phase advances of melatonin# body
temperature# or other circadian processes# suggests David Avery# a
psychiatrist at tha University of Washington School of Medicine in
Seattle. Low-lntensity lights turned on early in the morning advance
melatonin secretion in rats by a couple of hours# but no comparable
studies have been done in humans# Avery notes.
Zn 1990# Michael Terman reported the first pilot study of
slmulated-dawn treatments. After two weeks of awakening to synthetic
sunrises# six of eight SAD patients experienced total relief from
symptoms of depression.
Avery's team also reports encouraging results with dawn
simulation. Zn two studies, a total of 23 SAD patients spent one week
with a special incandescent bulb in their bedrooms that gradually
produced light comparable to a natural dawn between 4 and 6 a.m.
Another 18 SAD recruits received one week of placebo treatment with
either 30-minute dawns that peaked at a level comparable to moonlight or
two-hour dawns of slightly greater intensity. Symptoms of depression
largely cleared up only among those receiving full-scale dawn exposures.
Yet the ways in which bright light rewinds the human biological
clock remain unclear. For instance# a report in the January/February
N E U R O SC IE N C E L E T T E R S indicates that bright light does not reset an
important type of circadian rhythm in healthy adults in a pattern
consistent with what most SAD researchers had assumed.
Researchers led by physiologist David S. Minors of the University
of Manchester# England# administered 15-minute pulses of bright light to
15 college students at various times during the day or night. Exposure
to light significantly advanced the onset of daily increases in body
temperature when administered up to five hours after the participants'
average body temperature minimum# at around 5 a.m. investigators had
assumed that morning doses of bright light could spur phase advances In
SAD patients only if administered by 8 a.m.
On the other hand# doses of light in the evening did not delay the
drop in body temperature— another unexpected finding for researchers#
who had assumed that evening light would have such an influence on the
biological rhythms of SAD patients. Only very late— at 3 or 4 a.m.— did
bright light detain the circadian clock in the British study.
”For now# Z'd say that as long as you fail to got circadian phase
delay# light therapy probably works with SAD patients#" Michael Terman
maintains. ”But getting a phase advance with morning light therapy Is
not essential.”
The eye's sensitivity to light may play a key role in winter
depression# according to Terman. Preliminary studies in his laboratory
indicate that SAD individuals# compared with nondepressed controls,
report much more difficulty seeing a dim light after sitting for awhile
in a dark room# but only during the winter. Animal studies find that
the eye normally adapts to low levels of light during the day by
247
increasing the amount and sensitivity of light-absorbing pigments on the
retina, which transmits visual Information to tha brain.
SAD patients may have retinas incapable of squeeaing more light
out of shorter winter days, Terman suggests. Or the activity of their
retinal receptors that process daylight may plummet during tha winter in
a fashion that exaggerates decreases in retinal activity observed among
hibernating animals, he theorizes.
In any case, some aspect of simulated dawns other than modes light
intensity— perhaps thsir timing, gradual progression, or presentation at
particular times of ysar or latitudes— ‘serves to normalize light
sensitivity among winter depressives even while they sleep and their
eyelids are closes, Terman proposes.
Other studies, directed by Raymond H. Lam of the University of
British Columbia in Vancouver, add to the evidence of light
insensitivity during winter among SAD patients. One investigation,
involving 10 SAD patients and 19 nondepressed controls, measured
electrical activity across the retina in response to a short dose of
bright light* The retinas of persons with SAD generated weaker
electrical responses to light, Lam reports.
A second study measured electrical waves on the cornea that
respond in characteristic ways to light. Lam's group placed delicate
gold-foll electrodes on the anesthetized corneas of 24 SAD patients and
22 nondepressed volunteers. Overall, the two groups displayed the same
electrical-wave patterns following exposures to light. But the IB
seasonally depressed women generated smaller waves than tha 16 female
controls— in the former group— and the six seasonally depressed men
responded with larger waves than the six male controls.
Lam offers no explanation for this sex difference because he used
fairly crude measures of light sensitivity in a small sample.
Premenopausal women make up the large majority of SAD patients seen by
clinicians, but it remains unclear whether premenopausal women in
general stand a greater chance of suffering from the disorder.
Chemicals that carry messages from one brain cell to another, such
as dopamine and serotonin, are also present on the retina and take part
in communication between the eyes and tha brain, Lam notes. Disturbed
retinal function may interfere with visual processing medicated by these
chemical couriers.
"Surely the connections between eye and the brain are the key to
understanding seasonal affective disorder," says psychiatrist Norman B.
Rosenthal of the National Institute of Kental Health in Bethesda, Md.,
one of the pioneer SAD researchers.
Unfortunately, those connections start to dissolve when
researchers outfit SAD patients with "light visors" rather than sitting
them in front of light screens* The visors, only a few years ago
considered a promising treatment advance, encase two small incandescent
bulbs powered by a rechargeable battery. Study participants don a visor
and receive a 30-minute daily light dose for two weeks.
Two now studies suggest that light visors may function merely as
elaborate placebos, improving winter depression for reasons unrelated to
light exposure, Rosenthal asserts. In one project, directed by
psychiatrist Martin H. Teicher of HcLean Hospital in Belmont, Mass.—
which included Rosenthal on the research team— a din, red-light visor
designed to be ineffective reversed SAD symptoms slightly more than a
bright, white visor in a sample of 49 seasonally depressed patients.
And a study of 105 SAD patients in five Canadian and U.S. cities
indicated that dim, low intensity and high intensity light visors all
worked about equally well. About 40 percent of those receiving each of
the visor intensities recovered fully from symptoms of depression, and
60 percent in each group improved substantially, says Anthony J. Levitt
of the Clarke Insatiate of Psychiatry in Toronto.
248
"We don’t yet know what physical aspects of light have positive
effects on SAD patients," Levitt remarks. "That's disconcerting."
Michael Terman finds it more disconcerting that researchers have
not measured how much light from visors, screens, and simulated dawns
actually reaches the retina. With visors, light may focus on or above
the eyelid rather than on the eye itself, depending on the angle of the
headgear and the participant's direction of gase, Terman points out. He
argues that the low level of full recovery in Levitt's study, compared
with that in "light box" studies, suggests that volunteers wearing high-
intensity visors may actually have received the least light on their
retinas due to squinting, lowering their gase, or automatically
contracting their pupils in response to unavoidable bright light.
Moreover, the amount of light entering the retina during any type
of light therapy may fluctuate from one person to another, Terman
asserts. Tiny light meters attached to the foreheads of volunteers in
his laboratory reveal large individual differences in the amount of
light exposure from the same light screen.
"Light is a complex stimulus that has been inadequately specified,
given the intense clinical experimentation of tha last five years,"
Terman contends.
Mevertheless, light— -and the lack of it— -can really get under our
skin. For instance, rapid changes in day length greatly modify the
daily cycle of sleep and melatonin secretion, report researchers led by
psychiatrist Thomas A. Wehr of the National Institute of Mental Health.
Their findings with humans match those already observed in rats.
In Wahr's study, healthy adult volunteers spent one week in a
"summer" condition (16 hours of light per day, eight hours of sleep in a
dark room) and then healthy adult volunteers lived for four weeks in a
"winter" condition (10 hours of dally rest or sleep in a dark room.)
The shortened day resulted in participants sleeping more, the shortened
day resulted in participants sleeping more, reporting more sleepiness
while awake, experiencing a longer period of lowered body temperature
each night, and secreting more melatonin while asleep.
"Brain mechanisms that detect and respond to seasonal changes in
day length may have been conserved in the course of human evolution,"
Wehr suggests. "Artificial light may have profoundly modified patterns
of human sleep, temperature regulation, and hormone secretion."
Artificial light may also profoundly influence seasonal
depression, perhaps as much as shifts in day length among people who
rarely venture outside on winter days. People with full-blown and
milder forms of SAD often spend less time outdoors than Individuals
without seasonal problem, asserts Anna Wlrx-Justice, a psychologist at
the Psychiatric University Clinic in Basel, Switaerland. She finds that
SAD sufferers often improve as much by taking a daily one-hour walk in
normal winter sunlight as they do by soaking up a dally two-hour dose of
bright light indoors. The amount of sunshine available at a given
latitude may prove less important than the amount of time one actually
spends outdoors during the day, Wire-Justice proposes.
Ongoing studies of light treatments, the biological clock, and the
eye's sensitivity to light herald the dawn of a better understanding of
seasonal depression, Michael Terman contends. "We're at a point where a
new set of subtler questions about SAD and light treatment hae come to
the fore," he remarks.
Bower, B. (1992). Here comes the sun. Science News,
142(1) 142-143.
249
APPENDIX H
INSTRUCTIONS FOR THE READING TASK
Follow the instructions below. Please DO NOT unstaple the
packet. Bring the packet with you to the next class.
1. Read the attached article for next class.
2. Familiarise yourself with the ideas presented in the
article.
3. Feel free to underline, highlight, and make marks on the
article.
4. Feel free to make annotations in the margins of the
article.
5. During the next class, you will he asked to complete a
writing activity using the article.
6. NOW READ THE ARTICLE. In the space provided, indicate
the time you started reading the article.
TIME
250
APPENDIX I
SUMMARIZATION PROMPT
HEALTH SCIENCE ISO
STUDENT'S LAST NAME FIRST NAME M. I.
SCENARIOI
Fifty CSLA students who have not taken Health Science 150 are interested
in reading about some of the topics discussed in your claes. However,
because these students are participating in a special project, they do
not have enough time to read your Health Science ISO textbook.
Therefore, your instructor has agreed to have the studente read a
summary of the article you had to read for today.
Your job is to summarise the article The Safer Sex?
To summarise the article, you have 35 minutes.
INSTRUCTIONSI
As you write your eummary feel free tot
1. Look at the article*
2. Mark on the article if you wish.
3. Make changes. However, in making changes, do NOT erase your
original words, just draw a line through them.
4. Ask for more writing paper.
251
CULTURAL ANTHROPOLOGY 2SO
STUDENT'S LAST NAME FIRST NAME M. I.
SCENARIOI
Fifty CSLA students who have not taken Cultural Anthropology 250 are
interested in reading about some of the topics discussed in your class.
However, because these students are participating in a special project,
they do not have enough time to read your Cultural Anthropology 250
textbook. Therefore, your instructor has agreed to have the students
read a summary of the article you had to read for today.
Your job is to summarize the article Murder in Good Company
To summarize the article, you have 35 minutes.
INSTRUCTIONSl
As you write your summary feel free tot
1. Look at the article.
2. Hark on the article if you wish.
3. Hake changes. However, in making changes, do NOT erase your
original words, just draw a line through them.
4. Ask for more writing paper.
252
ANIMAL BIOLOGY 155
STUDENT'S LAST NAME FIRST NAME M. I.
SCENARIOI
Fifty CSLA students who havs not taken Animal Biology 155 are interested
in reading about some of the topics discussed in your class. However,
because these students are participating in a special project, they do
not have enough time to read your Animal Biology 155 textbook.
Therefore, your Instructor has agreed to have the students read a
summary of the article you had to read for today.
Your job is to summarize the article Here Comes the Sun
To summarize the article, you have 35 minutes.
INSTRUCTIONSI
As you write your summary feel free tot
1. Look at the article.
2. Mark on the article if you wish.
3. Make changes. However, in making changes, do NOT erase your
original words, just draw a line through them.
4. Ask for more writing paper.
253
APPENDIX J
PROPOSITIONS READ BY THE EXPERT READERS TO DETERMINE THE LEVEL OP
IMPORTANCE OP THE IDEAS
THE SAPBR SEX?
Probing a cardiac gandar gap
1. "Fiftaan yaara ago, 1.__
2. my mothar died quite prematurely of heart 2.__
disease.
3. Looking back, 3.__
4. There waa a failure to help her, 4.
5. from every corner, S.__
6. I have to feel that 6.__
7. aaid Conatance Horner, 7.__
8. under eecretary at the Department of Health and 8.__
Human Servicee,
9. addreaaing a 1989 conference 9.
10. on women and heart diaeaae. 10.
11. Horner*a atory offora a poignant reminder of a 11.__
traditional and often dangerous tendency*
12. to consider heart disease an affliction of men. 12.__
13. That assumption waa especially prevalent 13.
14. several decades ago, 14.
15. when physicians largely ignored women IS.
16. who complained of chest pain (angina), 16.__
17. a potential sign of impending heart attack. 17.
18. Heart disease represents the number-one killer of 18.__
men
19. in the United States, 19.
20. most people know. 20.
21. (heart disease) is also the number-one killer of 21.__
women,
22. but many don't realize that. 22.
23. Although the hormone estrogen apparently helps 23.
many women ward off coronary artery disease
24. during their reproductive years, 24.
25. this protective edge begins to vanish with 25.__
menopause.
26. Prom menopause on, 26.
27. a woman's risk of developing heart disease mounts 27.
rapidly,
28. soon equaling that of men. 28.
29. Statistics clearly bear this out* 29.__
30. Between age 45 and 65, 30.__
31. approximately one in nine U.S. women shows 31.
symptoms of heart disease,
32. compared with about one in seven men; 32.
33. after age 65, 33.
34. that ratio jumps to one in three, 34.
35. roughly matching the ratio in men. 35.__
36. But while the disease gap eventually disappears, 36.__
37. another gap lurks in the statistics* 37.__
38. there is a disquieting difference between men's 38.
and woman's cardiac mortality rates.
39. Women face twice the risk of death of men, 39.
40* during the first few weeks after a heart attack 40.
41. and their heightened relative risk persists well 41.
into the following year.
254
42. In addition, 42.
43. female candidates for coronary artery bypass
surgery must confront a stark facti
43.
44. the operation itself— 45--puts women in greater
jeopardy of death than men. * ■
44.
45. a risky but potentially lifesaving procedure 45.
46. Does this mortality gap reflect some innate,
biological difference in the female body?
46.
47. Are the statistical discrepancies due to known
risk factors,
47.
48. such as women's tendency to suffer heart attacks
at a later— 49— age than men
48.
49. and presumably frailer 49.
SO. To what extent do women or their physicians
overlook early symptoms of heart disease.
50.
51. delaying important treatment? 51.
52. Recently published research and findings— S3—
provide conflicting data and few answers.
52.
53. presented in November at the American Heart
Association (AHA) scientific sessions in Dallas
53.
54. Nevertheless, 54.
55. these collectively murky results are a first step
toward understanding the vagaries of heart
disease in women.
55.
56. according to many investigators. 56.
57. According to several research reports in the
1980s,
57.
58. women typically suffer attacks about a decade
later in life than men, and
58.
59. thus may not recover as readily as a younger
person of either sex.
59.
60. and this was the reason for women's higher death
rate in the 1980's.
60.
61. But the deadly difference doesn't spring solely
from age,
61.
62. new data from the University of Massachusetts
Kedical School in Worcester indicate.
62.
63. Cardiologist Richard C. Backer and his colleagues
studied 2,742 men and 597 women
63.
64. who were hospitalised for a first or second heart
attack.
64.
65. caused by a clot 65.
66. that restricted blood flow to the heart. 66.
67. within four hours after the attack. 67.
68. All patients received a clot-busting drug— 69—
within four hours after the attack.
68.
69. (recombinant tissue plasminogen activator) 69.
70. 9 percent of the women and 4 percent of the men
died
70.
71. during the initial hospital stay, 71.
72. Becker reported at the AHA meeting. 72.
73. And among those who survived the attack. 73.
74. 12 percent of the women died— 75— during the
following year compared with 6 percent of men
74.
75. (primarily from cardiovascular disease) 75.
76. The women in the study were more than twice as
likely as the men to develop chronic heart
failure
after the attack,
76.
77. 77,
78. the researchers found. 78.
79. In people with this potentially lethal condition, 79.
80. lingering damage from the acute attack 80.
permanently impairs the heart's pumping ability.
81. During the first few weeks after the heart 81.
attacks,
82* 5.7 percent of the women and 2.1 percent of the 82.
men developed chronic heart failure,
83. Becker reports. 83.
84. The patients' ages in this study followed a 84.
familiar pattern*
85. 24 percent of the women, and only 9 percent of 85.
the men were more than 70 years old.
86. And the age distribution did make a difference in 86.
the mortality risk,
87. Becker says B7.
88. but didn't account for the entire gender gap. 88.
89. In their statistical assessment of risk factors, 89.
90. the researchers also isolated preexisting, 90.
chronic conditions
91. that could complicate a heart patient's recovery. 91.
92. Compared with the men in the study, 92.
93. the women had a higher prevalence of high blood 93.
pressure and diabetes,
94. the researchers found. 94.
95. But even when the team statistically controlled 95.
for these conditions,
96. lumping them in with the age factor, 96.
97. a significant mortality gap remained. 97.
98. Becker can't explain the risk remnant, 98.
99. he says. 99.
100. But differences in female physiology might 100.
somehow tip the scales,
101. he and others speculate. 101.
102. "The heart of a woman doesn't tolerate a 102.
myocardial infarction
103 (heart attack) 103.
104. as well as that of a man," 104.
105. he suggests. 105.
106. However, 106.
107. another report argues against tha notion of a 107.
built-in risk for women.
108. Nicholas H. Fiebach and his colleagues at Vale 108.
University studied 332 female and 790 male heart
attack suffers,
109. finding that 14 percent of the women and 9 109*
percent of the men died
110. during the initial hospitalization 110.
111. When the researchers controlled not only for the 111.
risks associated with age, diabetes and
hypertension, but also for the severity of heart
disease,
112. they managed to account for all but 1,5 percent 112.
of the gender gap-
113. an insignificant remnant, 113.
114. Fiebach says. 114.
115. the severity of heart disease merits special 115.
attention in efforts to explain women's
heightened mortality risk,
116. the Yale researchers conclude. 116.
117. Women tend to reach the cardiac care unit 117.
lie. with more serious heart conditions than men. 118.
119. several of the findings suggest, 119.
120. they report in the Feb. 23, 1990 JOURNAL OF THE
AMERICAN MEDICAL ASSOCIATION.
120.
121. The women in this sample,— 122— , had more
severely impaired left ventricular function
121.
122. compared with the men, 122.
123. an important measure of heart attack severity. 123.
124. hospital admitting exams showed. 124.
125. Some physicians may overlook cardiac warning
signs
125.
126. in woman, 126.
127. especially if those patients are relatively
young,
127.
128. Fiebach believes. 128.
129. "If a middle-aged woman is having chest pain. 129.
130. ona doesn't think of a heart attack. 130.
131. If a middle-aged man is having chest pain. 131.
132. ona automatically thinks of a heart attack," 132.
133. he told science news. 133.
134. Heart disease may be more difficult to diagnose 134.
135. in women than In men, 135.
136. he also suggests. 136.
137. In any case, 137.
138. a failure to diagnose and treat coronary artery
disease— 139— could lead to more damaging heart
attacks,
138.
139. until it becomes more severe 139.
140. which in turn carry a higher risk of death, 140.
141. he says. 141.
142. The traditional diagnostic bias still influences
medical thinking,
142.
143. cardiovascular researcher, Kathleen B. Xing
agrees.
"There's definitely an attitude
143.
144. 144.
14S. that heart disease is a man's problem," 145.
146. says King at the University of Rochester in New
York.
146.
147. And physicians may not be alone 147.
148. in their delayed response to early warning signs,
or evsn an outright heart attack,
148.
149. in women. 149.
150. The women in his study gsnerally waited longer
than the men
150.
151. to head for tho hospital 151.
152. after their chest pains began, 152.
153. Fiebach found. 153.
154. "It may be that the longer you wait. 154.
155. the more chances are 155.
156. that you're going to have a severe complication," 156.
157. he suggests, 157.
158. while cautioning 158.
159. that the link between such delays and a poor
clinical outcome remains speculative.
159.
160. Physicians may treat male and female heart attack
patients differently
160.
161. once their stay in the hospital has ended. 161.
162. the Yale report also hints. 162.
163. Physicians prescribed beta-blocker drugs— 165— 163.___
more often for men than for women at the time of
hospital discharge
164. researchers noted. 164.
16S. which slow the heartbeat and may improve survival 165.___
outlook for some people.
166. the researchers noted. 166.
167. However, 167.
168. this finding did not reach statistical 168.___
significance in the study,
169. Fiebach emphasises. 169.
170. Further research is needed. 170.
171. to determine whether such gendsr-based treatment 171.
differences exist,
172. he says. 172.
173. The heart attack itself is only part of the 173.___
study.
174. Researchers also see disturbing differences
between the sexes
174.___
175. when they look at mortality rates during and
after coronary artery bypass surgsry,
175.___
176. in which surgeons attempt to restore blood flow 176.___
to tha heart by attaching a vein above and below
the site of a coronary artery blockage
177 to shunt blood around it. 177.
178. Whether undertaken before or after a heart 178.___
_.attack,
1797" bypass surgery can sometimes save lives 179.
180. but tha operation entails a greater gamble for
women than for men,
180.___
181. statistics show 181.
182. a research team led by cardiologist Steven S. 182.___
Khan at tho University of California, Los
Angeles, has uncovered evidence
183. suggesting that women with heart disease tend to 183.___
get bypass surgsry later in the course of their
diaeaae.
184. Khan attributes such delays— 185— to a "referral 184.___
bias" on the part of the physicians
185. which may increase the riskiness of the surgery. 185.
186. "Women pretty much have to have a heart attack or 186.
vary sever symptoms
187. life threatening symptoms of their heart disease 187.
188. before they are sent to surgery" 188.
189. Khan says. 189.
190. Some women 190.
191. who complain of chest pain get the brush-off by 191.___
physicians
192. who think the pain is psychosomatic, 192.
193. he suspects. 193.
194. like other researchers. 194.
195. Heart disease may be trickier to diagnose in
women,
195.___
196. whose reports of chest pain don't always match 196.___
the traditional description of angina,
197. he also notes. 197.
198. Horaovar, 198.
199. physicians may worry about advising some women to 199.
undergo the rigorous bypass procedure
200. unless their heart condition seems dire. 200.
258
201. Physicians nay worry about advising women 201.
202. debilitated by old age or 202.
203. suffering serious complicating condition. 203.
204. Khan and his colleagues assesses the mortality 204.
rates of 1,815 men and 482 women
205. who had bypass surgery between 1982 and 1987. 205.
206. 4.6 percent of the woman, 208 died 206.
207. during surgery or the postoperative recovery 207.
period
208. compared with 2.6 percent of the men. 208.
209. They report in the April 15, 1990 ANNALS OP 209.
INTERNAL HEDXCXNE.
210. But when the researchers statistically controlled 210.
for age and severity of illness,
211. the mortality gap vanished. 211.
212. The women were sicker than the men 212.
213. when wheeled into the operating room, 213.
214. showing more severe signs of heart disease, 214.
215. including a particularly risky type of chest pain 215.
216. called unstable angina. 216.
217. In general, 217.
218. they found. 218.
219. Regardless of gander, 219.
220. a patient's odds of dying— 221— increase— 222— 220.
221. during or after bypass surgery 221.
222. with the severity of his or her heart condition, 222.
223. says Khan. 223.
224. The notion— 225— finds no support in a study 224.
225. that physicians refer only the sickest women for 225.
surgery
226. conducted at Downstate Medical Center in 226.
Brooklyn, N.Y.
227. Thierry A. Folliguet and his co-workers assessed 227.
several measures of heart disease severity
228. including chest pain and the number of blocked 228.
arteries in 858 men and 439 women
229. who underwent bypass surgery at three New York 229.
hospitals in 1989.
230* Folliguet reported a total lack of evidence 230.
231. indicating that the women were more severely ill 231.
than the men
232. when they reached the operating room, 232.
233. at the AHA meeting. 233.
234.. He says 234.
235. he cannot explain the discrepancy between his 235.
finding and Khan's.
236. Folliguet favors a long-standing explanation of 236.
women's higher mortality rate from bypass
surgeryt
237. body else. 237.
238* Because women typically are smaller than men, 238.
239. their bypass operation pose more technical 239.
difficulties,
240. he says. 240.
Many cardiovascular surgeons espouse this view,
241. asserting that smaller arteries reduce a woman's 241.
chance of surviving the surgery and recovery
period.
242. Other researchers have uncovered data 242.
contradicting the body-size theory.
243. King focused on 464 women 243.
244. who had bypass surgery between 1983 and 1988# 244.
245. finding that 4.3 percent of them died 245.
246. during or eoon after the operation. 246.
247. A statistical analysis— 248— showed no link 247.
between bypass risk and body surface area
248. comparing the women who died with those who 248.
survived
249. at the AHA meeting, 249.
250. she reported. 250.
251. Body surface area may not correlate with the size 251.
of coronary arteries,
252. Ring quickly points out. 252.
253. Body surface is calculated from height and 253.
weight,
254. She and her colleagues want to conduct a follow- 254.
up study
255. measuring blood vessel size and tracking 255.
mortality rates,
256. in the hope of establishing whether or not small 256.
vessel size puts women— and smaller men— at extra
risk during bypass surgery.
257. The Rochester results also punch holes in a 257.
theory attributing women's higher bypass risk to
their tendency to suffer heart attacks at an
older age than men,
258. King's all-female sample showed no correlation 258.
between age and bypass mortality.
259. "Age really didn't make a difference," 259.
260. "our study showed,” 260.
261. she told SCIENCE NEWS. 261.
262. Potential complicating factors,— 263— similarly 262.
failed to explain these women's risk,
263. such as diabetes or high blood pressure, 263.
264. she says. 264.
265. A complex interplay of factors— 267— works 265.
against women
266. who suffer heart attacks 266.
267. including age, severity of illness and perhaps 267.
some innate biological differences,
268. many heart researchers today suspect. 268.
269. But that aquation remains largely unresolved. 269.
270. "He1re just beginning to explain how all these 270.
things contribute,"
271. Fiebach says. 271.
272. While no amount of research can save her mother 272.
now,
273. some 247,000 U.S. women face the prospect of a 273.
deadly heart attack in the coming year,
274. Horner points out. 274.
275. Given the magnitude and urgency of that threat, 275,
276. she urges researchers to pursue their 276.
investigations with vigor.
277. In the meantime, 277.
278. one fact ia cleart 278.
279. Men have no monopoly on hazards of the heart. 279.
MURDER XN OOOD COMPANY
Cooperation, camaraderie and a dizzying homicide rate distingulah
email New Guinea society
Proposition Proposition
Number Number
1. The Gebuei are a strikingly gentle lot. 1.
2. The Gebusi is a society of about 450 persons 2.
3. living in a New Guinea rain forest. 3.
4. They revel in kog-wa-yay, 4.
5. (which is) roughly translated as "good company." 5.
6. Togetherness, casual talk and exuberant humor are 6.
daily staples.
7. There is no central political structure and 7.
e. no jockeying for power among the stronger men; 8.
9. Matters of concern to Gebusi—>10— are decided by 9.
consensus.
10. who live in communal "longhouse" settlements 10.
11. Food—*12— is routinely shared among all the 11.
residents of a settlement.
12. including bananas grown in small gardens and the 12.___
occasionally hunted wild pig
13. Anger, violence and warfare are frowned upon. 13.
14. But behind this aura of serenity and conviviality 14.
lurks a brutal paradoxt
IS. The Oebusi murder one another at a rata among the 15.___
highest ever reported,
16. (The Gebusi murder rata is) about 40 times 16.___
greater than the 1960 homicide rate in the United
States.
17. Their murderous ways cannot be explained by 17.___
current theories of violence.
16. theories applied to societies that have complex 18.___
political and economic systems
19. such as ours, 19.
20. according to anthropologist Bruce M. Knauft of 20.
Emory University in Atlanta,
21. who documented the Gebusi homicide rate 21.
22. during nearly two yeare of field work, 22.
23. "The character of homicide appears to change in 23.
simple societies"
24. that have no pecking order or dominance hierarchy 24.___
among adult men.
25. says Knauft. 25.
26. "Especially in simple societies, 26.
27. there may be a pattern of social life that is 27.___
generally peaceful and tranquil but
26. (which) is punctuated by aggression 28.
29. which ie unrestrained and frequently homicidal. 29.
30. when it does occur 30.
31. This pattern may be a critical aspect of the
evolution of human violence
31.
32. since simple, decentralized societies have 32.__
predominated for most of the history of Homo
sapiens,
33. he adds. 33.
34. Anthropologists— 36— often come away with a 34.
vision of Eden in the outback,
35. "only to have the diecovery foiled by better
data"
35.
36. (anthropologists) who encounter euch societies 36.
37. says Emory anthropologist Melvin Konner. 37.
38. Margaret Mead's observations of life in Samoa are
perhaps the most famous instance of this
tendency.
38.
39. In 1928 39.
40. she wrote of a near-utopian Samoa 40.
41. unencumbered by aggression, competition, sexual
repression and conflict-ridden adolescence.
41.
42. Since 1950 42.
43. several accounts,— 45— have described a more
complex Samoa
43.
44. in which violent crime, including rape, is not
uncommon and
44.
45. is often committed by adolescent boys. 45.
46. including a controversial book by Australian
anthropologist Dersk Freeman
46.
47. In small societies of noncompetitive foragers or
hunter-gatherers,
47.
48. violence is especially easy to overlook, 48.
49. says Knauft. 49.
50. First, so.
51. ths people often fear and downplay occasional
aggressive outbursts, and
51.
52. second 52.
53. even a few murders committed from year to year
can translate into a high overall homicide rate.
S3.
54. A caee in point are the tKung Bushmen of the
Kalahari Desert,
54.
55. dubbed "the harmless people" 55.
56. by an Investigator several decades ago. 56.
57. They are indeed generally peaceful and
gregarious.
57.
58. but the IKung's homicide rate is nearly three
times that of the United States,
58.
59. which is already one of the highest among Western
nations
59.
60. a systematic survey in 1979 found. 60.
61. Within a population of 1,500 IKung, an estimated
22 killings occurred over five decades.
61.
62. "about five more than the same number of New
Yorkers would have been expected
62.
63. to commit over the same period," 63.
64. according to Konner. 64.
65. A similar pattern— 66— has been noted among
Central Eskimo groups, the Semai aborigines of
Malaysia and the nomadic Hadsa of Tansania.
65.
66. a pattern of pervasive good will and self-
effacement combined with occasional violent
flare-ups and even murders
66.
67. says Knauft. 67.
68. But the Gebusi study is perhaps the most
intensive homicide Inquiry to date.
68.
69. Knauft began by establishing complete genealogies
for 15 of 25 Gebusi clans and partial genealogies
for three others.
69.
70. Clan members are distantly or directly related to
one another.
70.
71
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
The total numbar of adult daaths in the
genealogical sample between 1940 and 1982 totaled
394.
The cauee of each death was cross-checked in
extensible discussions with Gebusi informants
relatives, friends and acquaintances of the
deceased person.
There were 129 cases of homicide, nearly one-
third of all deaths,
reports Knauft in the Aug-Oct. CURRENT
ANTHROPOLOGY.
The annual homicide rate from 1940 to 1982 is at
least 568 murders per 100,000 persons,
he says.
a conservative calculation based on the partial
population survey.
By contrast,
the 1980 homicide rate in the United States was
10.7 murders per 100,000;
according to the Federal Bureau of Investigation
statistics;
the estimate for Oetroit in 1985 was 58.2 murders
per 100,000 persons.
"Only the more extreme instances of modern mass
slaughter would equal or surpass the Gebusi
homicide rate over a period of several decades,"
says Knauft.
Four out of five Gebusi murders involved the
killing of someone branded as a sorcerer
(A sorcerer is accused of) having allegedly
caused the death of another Gebusi.
The murders were uncovered in the study
Death from diseases— 89— is ever-present among
the Gebusi and sets the stage for sorcery
accusations
often caused by infections and parasites
More than one-quarter of the sickness deaths in
the genealogical sample precipitated a sorcery
killing.
Although sorcery accusations are leveled at both
man and women,
only men carry out the homicides,
says Knauft
"including soma who were among the least
assertive and aggressive even by Gebusi
standards."
In two communities he surveyed,
11 of 17 adult men had killed an alleged
sorcerer.
a sorcerer conceals his or her evil intentions
toward others.
The Gebusi believe.
After a sickness death,
the firet step in uncovering a possible sorcerer
is to hold an all-night seance,
conducted by mediums
who believe their bodies are temporarily
inhabited by spirits.
If the seance provides a sorcery suspect,
that person must pass a "divination inquest."
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105. In many cases, 105.
106. the accused must cook a large piece of fish or
meat without burning it
106.
---
107. coated with powder ground from sago palms. 107.
108. Fires are placed around the food, rather than
under it.
108.
---
109. making proper cooking more difficult. 109.
110. Cooking is performed on the hearth of the dead
person.
110.
---
111. whose spirit is though to place its hands
underneath the food
111.
---
112. to ensure proper heating for Innocent sorcery
suspects.
112.
---
113. If the outcome is damning. 113.
114. the suspect may be killed on the spot, 114.
115. with no intervention by his or her kin. 115.
116. In some cases, 116.
117. close kin are ambivalent about a suspect's guilt, 117.
118. causing accusers to wait weeks or months 118.
119. before they organise a group 119.
120. the alleged sorcerer is ambushed in the forest
away from supporters,
120.
---
121. says Knauft. 121.
122. Revenge killings are rare. 122.
123. even after an ambush 123.
124. notes Knauft. 124.
125. Only four of the reported sorcery murders
resulted in a return killing by relatives of the
slain parson.
125.
126. Most killings in recent years have taken place
with the implied consent of all the accused
sorcerer's close kin
126.
127. implied consent includes mainly lack of vocal
support,
127.
128. Knauft points out. 128.
129. Otherwise, 129.
130. there is the possibility that a grieving relative
could report the murder to New Guinea government
authorities.
130.
131. A couple of cases have resulted in reports to
state officials and
131.
132. have resulted in imprisonment of the killers. 132.
133. Both cases involved the killing of an accused
sorcerer in a fit of rage.
133.
134. A few other killings in Knauft's sample occurred
during ritualistic bow-an-arrow skirmishes
between Gebusi from different settlements.
134.
135. Add it all up, and 135.
136. the peaceful Gebusi have quite a bit of blood on
their hands.
136.
137. But since sorcery is probably the last thing any
Gebusi of sound mind would practice,
137.
138. Why are so many murdered for their alleged
witchcraft?
138.
139. current theories of violence offer no
satisfactory answers,
139.
140. contends Knauft. 140.
141. For example. 141.
142. one characteristic of violence in human societies
is a kind of "genetic selection,"
142.
264
143. sociobiologiets have proposed. 143.
144. In the view of genetic selection, 144.
145. an offender is more likely to murder a
genetically unrelated victim
145.
146. in order to minimise the loss to the gene pool of
persons who share a high percentage of his or her
genes.
146.___
147. draws on the notion that there is an innate
tendency toward aggression.
147.___
148. Oebusl violence— 149— seems to defy this
prediction
148.___
149. however 149.
ISO. In a breakdown of genetic relationships in one
community of 30 men and 34 women.
ISO.
151. homicide victims were far more likely to be
killed by relatives living in the same longhouse
settlement,
151.___
152. Knauft found. 152.
153. Local groups of related males in simple societies
compete and fight with other "fraternal interest-
groups."
153.
154. another theory holds 154.
155. (This theory) springs from the position that
violence is learned.
155.___
156. These hostile bands are said to be precursors of
military organisations in larger, more complex
societies.
156.___
157. But aggressive, men-only organisations are
lacking in Gebusi society.
157.___
158. explains Knauft. 158.
159. In fact, 159.
160. married men often move to the settlements of
their in-laws,
160.___
161. creating overlapping social ties throughout the
society.
161.___
162. In general, 162.
163. men as well as women go out of their way to share
food and other resources.
163.___
164. Displaying dominance or superiority over other is
treated as a breach of "good company."
164.___
165. A third possibility is that 165.
166. an emphasis on punishment and obedience in child-
rearing— 167— promotes aggressive adult male
behavior.
166.
167. by authoritarian, emotionally distant fathers 167.
168. based on several cross-cultural studies 168.
169. Here again, 169.
170. the Gebusi are theory busters 170.
171. says Knauft. 171.
172. Physical punishment of children in their society
is extremely rare.
172.___
173. Fathers are affectionate and undemanding of their
sons,
173.___
174. he observes. 174.
175. "and in two years I only once saw a father so
much as shove his son in irritation."
175.___
176. The greatest threat to a child's thrust is the
stark realization that mother or father may soon
die.
176.___
26S
177.
178.
179.
180.
181.
182.
183.
184.
185.
186.
187.
188.
189.
190.
191.
192.
193.
194.
195.
196.
197.
198.
199.
200.
201.
202.
203.
204.
205.
206.
207.
208
209.
210.
either from natural causes or at the hands of
sorcery accusers.
According to Knauft,
in many cases "sorcery homicide is ultimately
about male control of marriageable women.”
Although Gebusi rarely discuss the balance of
women in a clan or smaller living groups,
there is an underlying expectation that— 182— , a
close female relative of their husband's will—
183— bo encouraged to marry back into the woman's
clan of origin,
after a woman marries
at some point
A strikingly high number of sorcery accusations
and murders are aimed at members of a group
thought to be holding out on its obligation to
promote marital exchange,
says Knauft.
It is not unusual for a woman to be accused of
sorcery and
killed by her own brothers
as a way of taking vengeance on their brothers-
in-law
he notes.
Nevertheless,
the Gebusi deny that there are male disputes over
women.
"The heavy cultural pressure against anger
expression seems to maximise the problems in
their system of marital exchange,"
says Knauft.
In simple societies such as the Gebusi,
marked by a lack of male status distinctions,
violence revolves around the control of sexual
relations and
sporadic, violent incidents serve to reestablish
cooperation in the community,
Knauft proposes.
A number of anthropologists welcome Knauft'e
extensive homicide data but
add their own caveats or interpretations to the
findings.
A history of being raided by more powerful
neighboring societies contributes to an attitude
of helplessness,— 203--, among people like the
IKung, Semai and Gebusi.
rather than harmlessness
Both Leland Donald of the University of Victoria,
British Columbia, and Robert K. Dentan of the
State University of New York in Buffalo suggest.
Resignation to the inevitability of periodic
massacres can encourage a cultural retreat from
angry and violent expressions,
Donald and Dentan suggest.
Semai communities are also being pressured by the
Halay governments
to give up their land for industrial development
and move to urban center
says Dentan,
who has done field work among the Semai.
177.
178.
179.
180.
181.
182.
183.
184.
185 .___
186.
187.
188.
189.
190.
191.
192 .___
193.
194.
195.
196.
197 .___
198 .___
199.
200 ._
201._
202.
203.
204.
205.
206.
207.
208 .___
209.
210.
266
211.
212.
213.
214.
215.
216.
217.
218.
219.
220.
221.
222.
223.
224.
225.
226.
227.
228.
229.
230.
231.
232.
233.
234.
235.
236.
237
238
239.
240.
241.
242.
243
244.
Although tensions ere building and
Semai murders are likely to increase,
Dentan has no evidence of periodic Semai homicide
related to male rivalry over marriageable women
"either of arrogant outsiders or of friends in
drunken outburst,"
"Knauft's account is probably the best
description of capital punishment in the
ethnographic literature,"
comments Keith F. Otterbein of the State
University of New York in Buffalo.
Gebusi killings are usually viewed as legitimate,
seldom involve retaliation and involve the
premeditated execution of a suspect found guilty,
elements that define capital punishment
Otterbein says.
Sudden unplanned killings occur more often among
the Eskimo and IKung,
a homicidal variation that Knauft does not
account for
Otterbein notes.
However,
the line between accepted and Inappropriate
killings in simple societies "can become
vanishingly thin"
Knauft says.
In the case of the Gebusi,
some sorcery murders are accepted by the
community
while others are actively opposed or tolerated
only out of cowardice or feat by the victim's
supporters.
There is also vanishingly thin line between
Knauft's explanation of Gebusi homicide and
socioblological theory,
write Hartin Daly and Margo Wilson— 233— in a
commentary accompanying Knauft's article
of HcMaster University in Hamilton, Ontario.
Pegging sorcery killings to male control of
marriageable women — 235— echoes the
socioblological emphasis on male competition—
236—
they contend.
to reproduce with available females as an
important spur to violence and homicide
"It is clear that homicides,— 238— , arise out of
conflicts of interest,"
among the Gebusi as among other peoples
according to Daly and Wilson.
They also question whether Knauft'e reliance on
the recollections of Gebusi individuals has
distorted his findings.
Knauft did not observe any homicides for himself.
Extensive cross-checking of each homicide
account,— 243— , produced a relatively accurate,
conservative estimate of the murder rate,
as well as the details of associated seances and
divinations
responds Knauft.
211.
212.
213.
214.
215.
216.
217.
218.
219.
220.
221
222.
223.
224.
225.
226.
227.
228.
229.
230.
231.
232.
233.
234.
235.
236.
237.
238.
239.
240.
241.
242.
243.
244.
245. On two occasions, 245
246. ho adds, 246
247. hit proaenco and that of his colleague and wife
Bileen M. Cantrell may have prevented a sorcery
killing.
247
248. In one case, 248
249. the anthropologists gave verbal support: to the
relatives of a woman
249
250. who was being threatened by her accusers. 250
251. It was a classic ethical challenge for those
studying homicide.
251
252. "He couldn't just stand by and let someone be
killed,"
252
253. says Knauft. 253
READER'S NAME
HERE COMES THE SUN
Scientists shad new light on winter dapraaaion
Proposition Rating
Number
1. A young woman lies asleep on an overcast winter
morning* 1. _
2. At 4 a.m., a faint low emanates from a light bulb
placed near her bed. 2.___
3. The glow gradually gains intensity and 3.___
4. (the glow) bathes the room in soft light by 6
a.m., 4.__
5. whan the woman awakens. 5.__
6. she has just welcomed the simulated dawn of a
new day. 6.___
7. After several mornings brightened with light 7.__
8. designed to mimic a bona fide sunrise, 8.__
9. the clouds begin to lift from the woman's
"winter depression," 9. _
10. (winter depression) which appears punctually in
late November of every year 10. _
11. and abates by the following April. 11. _
12. Preliminary support for the beneficial
"sleeper effect" of simulated dawns adds an
intriguing new twist to the much investigated
thaw of winter depression, 12. _
13. following one or two weeks of dally exposure
to bright lights. 13.
14. In fact, 14.__
15. as the study of seasonally recurring
depression enters its second decade, IS. __
16. researchers possess an abundance of illuminating
new findings, 16. _
17. that in true scientific fashion, raise complex
new questions. 17. _
18. Seasonally recurring depression is known by
the apt acronym SAD, 18. _
19. for seasonal affective disorder. 19.__
20. Much of the latest research was presented at the
recent annual meetings of the Society for Light
Treatment and Biological Rhythms in Bethesda,
Md. and the American Psychiatric Association
in Washington, DC. 20. _
21. Along with feelings of sadness, anxiety, and
lethargy, people experiencing SAD often display
symptoms 21. __
22. not typically seen in cases of nonseasonal
depression. 22.__
23. These include difficulty awakening in the
morning, daytime drowsiness, cravings for sweat
or starchy carbohydrate foods, and extreme weight
gain. 23.__
24. Social withdrawal and a marked drop in work
performance also characterise SAD, 24.__
25. which usually occurs for the first time in
people in their early 20s. 25. __
26. A large segment of the population suffers from
269
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
winter depression, 26.
with the numbers increasing in regione farther
from the equator, 27.
where winter nights grow longest (SNt9/23/89,
p. 198). 28.
Several independent studies estimate that the
condition affects 1 to 2 percent of Florida’s
population, 29.
about 6 percent of people living in Maryland
and New York City, and 30.
nearly 10 percent of the residents of New
Hampshire and Alaska. 31.
The prevalence of mild SAD symptoms rangss
from nearly 3 percent in Florida to 11 percent
in New Hampshire. 32.
About 15 percent of those hospitalised for severe
depression display a regular worsening or
reappearance of symptoms in the winter, 33.
according to studies conducted in the United
States and Europe. 34.
However, 35.
many SAD sufferers nsver seek mental health care,36.
apparently deciding to slog through each winter
in anticipation of a corrective dose of spring
sunshine. 37.
But numerous studies promote a faster
SAD-bustlng technique! 38.
The faster SAD-busting technique consists of
daily exposure to bright lights for a week or
two during the winter. 39.
The studies were conducted over the past decade 40
Researchers currently favor two approaches
to this treatment! 41.
The approaches include either placing an
individual in front of a screen emitting light
five times brighter than ordinary room light for
two hours each day,
or using a screen producing 42.
light 20 times brighter than normal room light
for 30 minutes each day. 43.
Morning sessions often yield the best results,
but 44.
evening exposures can also provide relief. 45.
Although the proportion of SAD patients— 47—
varles widely from study to study, 46.
whose symptoms disappear in response to
bright-light 47.
the majority of reports cite strong improvement
in at least 60 percent of participants, 48.
assert Michael Terman and Jluan Su Terman,
psychologists at the New York State Psychiatric
Institute in New York City, 49.
in a January 1991 review prepared for the
federal government's "depression guidelines
panel." 50.
A much smaller proportion of seasonally depressed
people feel markedly better SI.
after placebo treatment with dim lights, 52.
the researchers add. 53.
Observations of changes in the daily biological
270
ss.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75
76.
77.
78.
79.
80.
81.
82
83.
84.
85.
clock, or circadian rhythms, of rata--55--
inspirad tha first attempts to treat winter
depression with lights 54.
exposed to bright light at specific times 55.
Scientists have no uncontested theory to explain
how bright lights help human SAD sufferers, but 56.
(most scientists) assume that the treatment must
somehow alter circadian rhythms. 57.
SAD often derives, at least in part, from
the delayed nightly secretion of melanin, 58.
(melanin is) a hormone involved in tha regulation
of sleep; 59.
One theory holde 60.
Successful bright-light therapy can initiate
melatonin secretion an hour or two earlier, 61.
soma evidence suggests. 62.
Scientists refer to this readjustment of the
biological clock as a "phase advance." 63.
Simulated dawns may also prompt phase advances
of melatonin, body temperature, or other
circadian processes, 64.
suggests David Avery, a psychiatrist at the
University of Washington School of Hedicine in
Seattle. 65.
Low-intensity lights turned on early in the
morning advance melatonin secretion in rats by
a couple of hours, but 66.
no comparable studies have been done in humans, 67.
Avsry notes. 68.
In 1990, 69.
Michael Terman reported the first pilot study
of simulated-dawn treatments. 70.
After two weeks of awakening to synthetic
sunrises, 71.
six of eight SAD patients experienced
total relief from symptoms of depression. 72.
Avery's team also reports encouraging results
with dawn simulation. 73.
In two studies, 74.
a total of 23 SAD patients spent one week with
a special incandescent bulb in their bedrooms 75.
that gradually produced light comparable to a
natural dawn between 4 and 6 a.m. 76.
Another 18 SAD recruits received one week of
placebo treatment with either 30-mlnute dawns
that peaked at a level comparable to moonlight
or 77.
two-hour dawns of slightly greater intensity. 78.
Symptoms of depression largely cleared up only
among those receiving full-scale dawn exposures. 79.
Yet the ways in which bright light rewinds the
human biological clock remain unclear. 80.
For instance, 81.
bright light does not reset an important
type of circadian rhythm in healthy adults 82.
in a pattern consistent with what moat
SAD researchers had assumed, 83.
A report in the January/February NEUROSCIENCE
LETTERS indicates. 84.
Researchers administered 15-mlnute pulses of
bright light to 15 college students at various
times during the day or night* 65*
86. (Researchers were) led by physiologist David 8.
Minors of the University of Manchester, England. 86.
87. Exposure to light significantly advanced the
onset of daily increases in body temperature 87.
88. when administered up to five hours after the
participants' average body temperature minimum,
at around S a.m. 88.
89. morning doses of bright light could spur phase
advances in SAD patients only if administered
by 8 a.m., 89.
90. Investigators had assumed. 90.
91. On the other hand, 91.
92. doses of light in the evening did not delay
the drop in body temperature— 92.
93. another unexpected finding for researchers, 93.
94. who had assumed that evening light would have
such an influence on the biological rhythms of
SAD patients. 94.
95. Only very late— at 3 or 4 a.m. did bright light
detain the circadian clock in the British study. 95.
96. "For now, I'd say that as long as you fail to
get circadian phase delay, light therapy probably
works with SAD patients," 96.
97. Michael Terman maintains. 97.
98. "But getting a phase advance with morning light
therapy is not essential." 98.
99. The eye's sensitivity to light may play a
key role in winter depression, 99.
100. according to Terman. 100.
101. SAD individuals report much more difficulty
seeing a dim light after sitting for awhile
in a dark room, 101.
102. compared with nondepressed controls 102.
103. but only during the winter, 103.
104. preliminary studies in his (Terman*s) laboratory
indicate. 104.
105. The eye normally adapts to low levels of light
during the day 105.
106. by increasing the amount and sensitivity of
light-absorbing pigments on the retina, 106.
107. which transmits visual information to the brain,107.
108. animal studies find. 108.
109. SAD patients may have retinas incapable of
squeesing more light out of shorter winter days,109.
110. Terman suggests. 110.
111. Or the activity of their retinal receptors that
process daylight may plummet during the winter 111.
112. in a fashion that exaggerates decreases in
retinal activity observed among hibernating
animals, 112.
113. he theorises. 113.
114. In any case, 114.
115. some aspect of simulated dawns serves to
normalise light sensitivity among winter
depressives 115.
116. other than modest light intensity 116.
117. even while they sleep and their eyelids are
closed, 117.
272
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
141.
142.
143.
144.
145.
Some aspects of simulated dawns other than
modest light include their timing* gradual
progression, or presentation at particular
times of year or latitudes 118.
Terman proposes. 119.
Other studies add to the evidence of light
insensitivity during winter among SAD patients 120.
(studies) directed by Raymond W. Lam of the
University of British Columbia in Vancouver. 121.
One investigation,— 123— measured
electrical activity across the retina in
response to a short dose of bright light. 122.
involving 10 8AD patients and 19 nondepressed
controls 123.
The retinas of persons with SAD generated
weaker electrical responses to light, 124.
Lam reports. 125.
A second study measured electrical waves on the
cornea that respond in characteristic ways to
light. 126.
Lam's group placed delicate gold-foll electrodes
on the anesthetieed corneas of 24 SAD patients
and 22 nondepressed volunteers. 127.
Overall, 128.
the two groups displayed the same
electrical-wave patterns following exposures
to light. 129.
But the 18 seasonally depressed woman generated
smaller waves than the 16 female controls 130.
indicating weaker light sensitivity in the
former group— 131.
and the six seasonally depressed men responded
with larger waves than the six male controls. 132.
Lam offers no explanation for this sex
difference 133.
because he used fairly crude measures of light
sensitivity in a small sample. 134.
Premenopausal women make up the large majority
of SAD patients seen by clinicians, but 135.
it remains unclear whether premenopausal women
in general stand a greater chance of suffering
from the disorder. 136.
Chemicals that carry massages from one brain
cell to another— 138— are also present
on the retina and 137.
such as dopamine and serotonin 138.
take part in communication between the eyes and
the brain, 139.
Lam notes. 140.
Disturbed retinal function may interfere with
visual processing medicated by these chemical
couriers. 141.
"Surely the connections between eye and the
brain are the key to understanding seasonal
affective disorder," 142.
says psychiatrist Norman B. Rosenthal of the
National Institute of Mental Health in
Bethesda, Md., 143.
one of the pioneer SAD researchers. 144.
Unfortunately, 145.
146.
147.
148.
149.
150.
151.
152.
153.
154.
155.
156.
157.
158.
159.
160.
161.
162.
163.
164.
165.
166.
167.
168.
169.
170.
171.
172.
173.
174.
175.
176.
177.
those connections start to dissolve 156.
when researchers outfit SAD patients with
"light visors” rather than sitting them in
front of light screens. 147.
The visors,--149— encase two small
incandescent bulbs powered by a rechargeable
battery. 148.
only a few years ago considered a promising
treatment advance, 149.
Study participants don a visor 150.
and receive a 30-minute daily light dose for
two weeks. 151.
Light visors may function merely as elaborate
placebos, 152.
improving winter depression for reasons
unrelated to light exposure, 153.
two new studies suggest, 154.
Rosenthal asserts. 155.
In one project, 156.
a din, red-light visor designed to be
ineffective reversed SAD symptoms slightly
more than a bright, white visor 157.
in a sample of 49 seasonally depressed patients.158.
(The project was) directed by psychiatrist
Martin H. Teicher of McLean Hospital in Balmont,
Mass, 159*
(the project) which included Rosenthal
on the research team. 160.
And dim, low intensity and high intensity
light visors all worked about equally well 161.
a study of 105 SAD patients indicated. 162.
(The study was conducted) in five Canadian
and U.S. cities. 163.
About 40 percent of those receiving each of tha
visor intensities recovered fully from symptoms
of depression, and 164.
60 percent in each group improved substantially,165.
says Anthony J. Levitt of the Clarke Institute of
Psychiatry in Toronto. 166.
”We don't yet know what physical aspects of
light have positive effects on SAD patients," 167.
Levitt remarks. 168.
"That's disconcerting.” 169.
It is more disconcerting that researchers have
not measured how much light from visors, screens,
and simulated dawns actually reaches the retina.170.
Michael Terman says. 171.
With visors, light may focus on or above the
eyelid rather than on the eye itself, 172.
depending on the angle of the headgear and
tha participant's direction of gaze, 173.
Terman points out. 174.
The low level of full recovery in Levitt's
study, — 178— suggests that volunteers wearing
high-intensity visors may actually have
received the least light on their retinas 175.
when compared with that in "light box" studies, 176.
due to squinting, lowering their gaze, or
automatically contracting their pupils in
response to unavoidable bright light* 177.
178.
179.
180.
181.
182.
183.
184.
185.
186.
187.
188.
189.
190.
191.
192.
193.
194.
195.
196.
197.
198.
199.
200.
201.
202.
203.
204.
205.
206.
He (Terman) argues.
Moreover, the amount of light entering the
retina during any type of light therapy may
fluctuate from one person to another,
Terman asserts.
Tiny light meters attached to the foreheads
of volunteers in his laboratory reveal large
individual differences
in the amount of light exposure from the same
light screen.
"Light is a complex stimulus that has been
inadequately specified,
given the intense clinical experimentation of
the last five years,*
Terman contends.
Nevertheless,
light— and the lack of it— can really get
under our skin.
For instance,
rapid changes in day length greatly modify
the daily cycle of sleep and melatonin
secretion,
report researchers led by psychiatrist Thomas
A. Wehr of tha National Institute of Mantal
Health.
Their findings with humans match those already
observed in rats.
In Hehr's study,
healthy adult volunteers spent one week in a
"summer" condition and
(The summer condition involved) 16 hours of
light per day, eight houre of sleep in a dark
room
then (healthy adult volunteers) lived for four
weeks in a "winter" condition
(The winter condition involved) 10 hours of
dally rest or sleep in a dark room
The shortened day resulted in participants
sleeping more, reporting more sleepiness while
awake, experiencing a longer period of lowered
body temperature each night, and secreting
more melatonin while asleep.
"Brain mechanisms that detect and respond
to seasonal changes in day length may have
been conserved in the course of human evolution
Wehr suggests.
"Artificial light may have profoundly modified
patterns of human sleep, temperature regulation
and hormone secretion."
Artificial light may also profoundly influence
seasonal depression,
perhaps as much as shifts in day length among
people
(people) who rarely venture outside on winter
days.
People with full-blown and milder forms of SAD
often spend less time outdoors than individuals
without seasonal problem,
asserts Anna Wirs-Juetice,
a psychologist at the Psychiatric University
178.
179.
180.
181.
182.
183.
184.
185.
186.
187.
188.
189.
190.
191.
192.
193.
194.
195.
196.
197.
198.
199.
200.
201.
202.
203.
204.
205.
207.
208.
209.
210.
211.
213.
214.
215.
216.
217.
218.
219.
Clinic in Basel, Switzerland.
SAD sufferers often improve as much
by taking a daily one hour walk in normal
winter sunlight
as they do by soaking up a daily two-hour
dose of bright light indoors,
she (Anna Wirz-Justice) finds.
The amount of sunshine— 213— may prove
less important than the amount of time
one actually spends outdoors
available at a given latitude
during the day,
Hirs-Justice proposes.
Ongoing studies of light treatments, the
biological clock, and the eye's sensitivity
to light herald the dawn of a better
understanding of seasonal depression,
Michael Terman contends.
"He*re at a point where a new set of subtler
questions about SAD and light treatment has
come to the fore,"
he remarks.
206.
207.
208.
209.
210.
212.
213.
214
215.
216.
217.
218.
219.
APPENDIX K
MAIN IDEAS IDENTIFIED IN THE ARTICLES
The Safer Sex
1. There ie a tendency to consider heart disease an affliction of men.
2. Heart disease represents the number one killer of men in the United
States.
3. But it is also tha number-one killer of women.
4. From menopause on, a woman's risk of developing heart disease
mounts rapidly, soon equaling that of men.
5. There is a disquieting difference between men's and women's cardiac
mortality rates.
6. Women face twice the risk of death of men during the first weeks
after a heart attack and their heightened relative riek persiete
well into the following year.
7. The operation puts woman in greater jeopardy of death than men.
8. Women typically suffer attacks about a decade later in life than
men.
9. The woman in a study were more than twice as likely as the men to
develop chronic heart failure after the attack.
10. Woman tend to reach the cardiac care unit with more serious heart
conditions than men.
11. Some physicians may overlook cardiac warning signs in women.
12. Heart disease may be more difficult to diagnose in women than in
men.
13. The traditional diagnostic bias still influences medical thinking.
14. Women in a study generally waited longer than the men to head for
the hospital after their chest pain began.
15. Physicians may treat male and female heart attack patients
differently once their stay in the hospital has ended.
16. There are differences in mortality rates during and after coronary
artery bypass surgery.
277
17. Homan with haart disease tend to get bypass surgery later in the
course of their disease.
18. Because women are smaller than men, their bypass operations pose
more technical difficulties.
19. A complex interplay of factors— including age, severity of
illness and perhaps some innate biological differences— works
against women who suffer heart attacks.
278
Murder In Oood Company
1* tn Gebusi society, anger, violence, and warfare are frown upon.
2. The Gebusi murder one another at a rate among the highest ever
reported.
3. The Gebusi murderous ways cannot be explained by current theories
of violence.
4. The character of homicide appears to change in simple societies.
5. Pour out of five Gebusi murders involve the killing of someone
branded as a sorcerer.
6. Current theories of violence offer no satisfactory answer to these
deaths.
7. One characteristic of violence in human eocieties is a kind of
genstic selection.
8. In the view of genetic selection, an offender is more likely to
murder a genetically unrelated victim.
9. Gebusi violence seems to defy the prediction that there is an
innate tendency toward aggression.
10. Gebusi homicide victims are more likely to be killed by relatives
living in the same longhouse settlement.
11. Another theory holds that local groups of related males in simple
societies compete and fight with other "fraternal-interest groups."
12. But aggressive, men-only organizations are lacking in Gebusi
society.
13. A third possibility is that an emphasis on punishmsnt and obedience
in child-rearing by authoritarian, emotionally distant fathers
promotes aggressive adult male behavior.
14. But physical punishment of children in their society is extremely
rare.
15. In many cases, sorcery homicide is ultimately about male control of
marriageable women.
16. A strikingly high number of sorcery accusations and murders are
aimed at members of a group thought to be holding out on its
obligation to promote marital exchange.
17. Sporadic, violent incidents serve to reestablish cooperation in the
community.
279
Here_Coraes-the.Sun
1. Preliminary support for tha beneficial "sleeper effect" of
simulated dawns adds an intriguing new twist to the much
investigated thaw of winter depression*
2. Seasonally recurring depression is known by the apt acronym SAD,
for seasonally affective disorder.
3* Along with feelings of sadness, anxiety, and lethargy, people
experiencing SAD often display symptoms not typically seen in
cases of nonseasonal depression.
4. A large segment of the population suffers from winter depression,
with the numbers increasing in regions farther from the equator.
5. A faster SAD-busting technique consists of daily exposure to bright
lights for a week or two during the winter.
6. The majority of reports cite strong improvement in at least 60
percent of participants.
7. Tha treatment must somehow alter circadian rhythms.
6. Yet tha ways in which bright light rewinds the human biological
clock remain unclear.
9. The eye's sensitivity to light may play a key role in winter
depression.
10. Some aspect of simulated dawns other than modest light intensity
serves to normalize light sensitivity among winter depresaivea even
while they sleep and their eyelids are closed.
11. Artificial light may have profoundly modified patterns of human
sleep, temperature regulation, and hormone secretion.
12. Artificial light may also profoundly influence seasonal depression,
perhaps as much as shifts in day length among people.
13. People with full-blown and milder forma of SAD often spend less
time outdoors than individuals without seasonal problem.
14. Ongoing studies of light treatments, the biological clock, and the
eye's sensitivity to light herald the dawn of a better
understanding of seasonal depression.
280
APPENDIX L
Tha
1.
2 .
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
IDEA UNITS IDENTIFIED IN THE ARTICLES
Safer Sex
There is a tendency
to consider heart disease an affliction of men.
Heart disease represents the number one killer of men
in tha United States.
It is also the number-one killer of women.
From menopause on,
a woman’s risk of developing heart disease mounts rapidly,
soon equaling that of men.
[There is] a disquieting difference between men's and women's
cardiac mortality rates.
Women face twice the risk of death of men
during the first weeks after a heart attack
and their heightened relative risk persists
well into the following year.
The operation puts women in greater jeopardy of death than men.
Women typically suffer attacks
about a decade later in life than men.
The women in a study were more than twice as likely as the men
to develop chronic heart failure
after the attack.
Women tend
to reach tha cardiac care unit
with more serious heart conditions than men.
Some physicians may overlook cardiac warning signs
in women.
Heart disease may be more difficult
to diagnose in women than in men.
The traditional diagnostic bias still influences medical thinking.
Women in a study generally waited longer than the men
to head for the hospital
after their chest pain began.
Physicians may treat male and female heart attack patients
differently
once their stay in the hospital has ended.
There are differences in mortality rates
during and after coronary artery bypass surgery.
Women with heart disease tend
to get bypass surgery
later in the course of their disease.
Because women are smaller than men,
their bypass operations pose more technical difficulties.
A complex interplay of factors— 42— works against women
who suffer heart attacks.
including age, severity of illness and perhaps some innate
biological differences
281
Murder In.Oood Company
1. In Oebuei society,
2. anger, violence, and warfare are frown upon*
3. The Oebuei murder one another
4. at a rate among the highest ever reported.
5. The Oebusi murderous ways cannot be explained by current theories
of violence.
6. The character of homicide appears
7. to change in simple societies.
8. Pour out of five oebusi murders involve the killing of someone
9. branded as a sorcerer.
10. Current theories of violence offer no satisfactory answer
11. to these deaths.
12. One characteristic of violence— 13— is a kind of genetic selection.
13. in human societies
14. in the view of genetic selection,
15. an offender is more likely
16. to murder a genetically unrelated victim.
17. Oebusi violence seems
18. to defy the prediction that
19. there is an innate tendency toward aggression.
20. Oebusi homicide victims are more likely
21. to be killed by relatives
22. living in the same longhouse settlement.
23. — 25— local groups of related males— 24— compete and fight with
other "fraternal-interest groups."
24. in simple societies
25. Another theory holds that
26. But
27. aggressive, men-only organisations are lacking
28. in Oebusi society.
29. an emphasis on punishment and obedience— 30— by authoritarian,
emotionally distant fathers promotes aggressive adult male
behavior.
30. in child-rearing
31. A third possibility is that— 29—
33. But
33. physical punishment of children— 34— is extremely rare
34. in their society.
35. In many cases,
36. sorcery homicide is ultimately about male control of
marriageable women.
37. A strikingly high number of sorcery accusations and murders
are aimed at members of a group
38. thought to be holding out
39. on its obligation to promote marital exchange.
40. Sporadic, violent incidents serve
41. to reestablish cooperation in the community.
282
Here Cornea the Sun
1.
2.
3.
4.
5 .
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
Preliminary support for the beneficial "Bleeper effect” of
simulated dawns adds an intriguing new twist to the much
investigated thaw of winter depression.
Seasonally recurring depression is known by the apt acronym SAD,
for seasonally affective disorder.
Along with feelings of sadness, anxiety, and lethargy,
people experiencing SAD often display symptoms
not typically seen in cases of nonseasonal depression.
A large segment of the population suffers from winter depression,
with the numbers increasing in regions farther from the equator.
A faster SAD-busting technique consists of daily exposure to bright
lights
for a week or two
during the winter.
The majority of reports cite strong improvement
in at least 60 percent of participants.
Tha treatment must somehow alter circadian rhythms.
Yet
tha ways— 17— remain unclear.
in which bright light rewinds the human biological clock
The eye's sensitivity to light may play a key role
in winter depression.
Some aspect of simulated dawns— 21— •serves— 22-
other than modest light intensity
to normalise light sensitivity among winter depressives
even while they sleep
and their eyelids are closed.
Artificial light may have profoundly modified patterns of human
sleep, temperature regulation, and hormone secretion.
Artificial light may also profoundly influence seasonal depression,
perhaps as much as shifts in day length
among people.
People with full-blown and milder forms of SAD often spend less
time outdoors than individuals without seasonal problems.
Ongoing studies of light treatments, the biological clock, and the
eye's sensitivity to light herald the dawn of a
better understanding of seasonal depression.
283
APPENDIX M
HOLISTIC RATING SCALE FOR SUMMARY WRITING
Scores
6. Demonstrates clear competence in summary writing at the content,
rhetorical, and syntactic levels, though it may hant occasional errors.
A summary in this category*
-provides the gist of the article (through macropropositions) or
contains all of the major propositions
-represents the emphasis and same pragmatic intent of the original
text
-is well organized and well developed
-displays consistent facility in the use of language and syntax,
-includes accurate paraphrases from the original text
-displays consistent objectivity
-is conceptually restricted to the content of the original text
S. Demonstrates competsnce in summary writing at—ths content,
rhetorical, and syntactic levels, though it will probably have
occasional errors
A summary in this category*
-contains most of the major propositions in the original text
-represents a similar emphasis and pragmatic intent as that in the
original text
-may be conceptually broader than the original text
-is generally well organized and well developed
-displays facility in the use of language and syntax
-includes paraphrases from the original text, though
Inconsistently
-displays objectivity
4. Demonstrates minimal competsnce in summary writing at the content,
rhetorical and syntactic levels
A summary in this category*
-contains some of the major propositions of the original text
-may contain some errors in representing the emphasis and
pragmatic intent of the original text
-is adequately organized and developed
-presents adequate but possibly inconsistent facility with
syntax and usage
-may contain some errors that occasionally obscure meaning
-may include some near copies of the original text
-may contain comments and/or observations about the original text
3 Demonstrates some developing competence in summary writing, but
remains flawed at either the content, rhetorical, or syntactic level
A summary in this category may reveal one or more of the following
weaknesses*
-fails to represent the main propositions in the original text
-fails to represent the emphasis and pragmatic intent of
the author
-fails to adequately organize or develop the ideas
-presents unnecessary information
-pressnts errors in sentence structure and inappropriate choice of
words
-pressnts some copies from the original text
-presents comments and/or observations about the original text
284
2. Suggests incompetence in susury writing.
A summary in this category is seriously flawed by one or more of
the following weaknesses!
-fails to represent the propositions of the original text
-fails to demonstrate an understanding of the author's emphasis
and pragmatic intent
-is seriously disorganised or underdeveloped
-includes irrelevant information
-hae serious and frequent errors in sentence structure or
usage
-frequently copies/quotes the original text
-frequently includes the reader/writer's comments and/or
observations about the original text
1. Demonstrates incompetence in sumury writing.
A summary in this category!
-may be incoherent
-may be underdeveloped at the content level
-may contain severe and persistent writing errors
-may copy the original text
-may consistently Include comments and general observations
about the original text
Adapted from The TOEFL test of written English guide. (19B9). By Snow,
H. A., Xamhi-Stein, L. D., £ Dwyer, A.
285

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

A Comparison Of Selected Writing Criteria Used To Evaluate Nonnative Speakers Of English At A California State University

PDF

The Nature Of Assisted Performance With Learning Handicapped Students During Language Arts

PDF

Self-selected reading and interactive vocabulary instruction: knowledge and perceptions of word learning among L2 learners

PDF

Communicative rights and responsibilities in an East Los Angeles barrio: An analysis of epistemic modal use

PDF

A comparative analysis of factors underlying the reading performance of American and South African ESL university students

PDF

Effective practices of a Spanish -bilingual and -bicultural principal vs. a non -Spanish -bilingual and -bicultural principal and the effects of their language proficiency and cultural knowledge ...

PDF

Identity politics and accessing discourses: SLA narratives of Korean immigrants

PDF

The Effect Of Efficacy-Building Instruction On Cognitive Engagement And Achievement

PDF

Coordinating the reading and writing competencies during the early literacy development of 5--6 year -old Spanish -English emergent bilinguals

PDF

The Effects Of Music Field Trips In A High School Music Appreciation Curriculum On Cognitive Music Learning, Including Language Acquisition And Music Attitudes

PDF

Correlation of factors related to writing behaviors and student -developed rubrics on writing performance and pedagogy in ninth-grade students

PDF

The Effects Of Selected Variables On The Transitioning Of Low-Achieving Special Education Adolescents From High School

PDF

Week of the Crystal Night: An original screenplay

PDF

A multicultural curriculum's relationship to attitudes of prejudice and stereotyping in 5th and 6th-grade students

PDF

Implementing and assessing problem -based learning in non -traditional post -secondary aviation safety curricula: A case study

PDF

Improving the academic achievement of African-American males: a case study in San Diego, California

PDF

An investigation of shame and anxiety in learning English as a second language

PDF

Impact of Proposition 227 on bilingual teachers' beliefs and practices

PDF

An Assessment Of The Zooreach! Program As A Model For The Development Of Informal Education Programs

PDF

Access factors and homerun experiences: Making readers of our children

Asset Metadata

Creator Kamhi-Stein, Lia Diana (author)

Core Title The Effect Of Explicit Instruction On The Summarization Strategies Of "Underprepared" Native Spanish-Speaking Freshmen In University-Level Adjunct Courses

Contributor Digitized by ProQuest (provenance)

Degree Doctor of Philosophy

Degree Program Education

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag education, bilingual and multicultural,education, curriculum and instruction,education, language and literature,education, reading,OAI-PMH Harvest

Language English

Advisor Eskey, David E. (committee chair), Kaplan, Robert B. (committee member), Rueda, Robert (committee member), Snow, Marguerite Ann (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c20-590681

Unique identifier UC11226800

Identifier 9614032.pdf (filename),usctheses-c20-590681 (legacy record id)

Legacy Identifier 9614032.pdf

Dmrecord 590681

Document Type Dissertation

Rights Kamhi-Stein, Lia Diana

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA