University of Southern California Dissertations and Theses

The effects of intellectual functioning and mediation on conceptual learning for Black college students

(USC Thesis Other)

The effects of intellectual functioning and mediation on conceptual learning for Black college students

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content THE EFFECTS OF INTELLECTUAL FUNCTIONING AND MEDIATION
ON CONCEPTUAL LEARNING FOR BLACK COLLEGE STUDENTS
by
Felicia Friendly
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Réquirements for the Degree
DOCTOR OF PHILOSOPHY
(Psychology)
May 19 70
UMI Number: DP71535
All rights reserved
INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
in the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
UMI
Dissertation R jblishing
UMI DP71535
Published by ProQuest LLC (2015). Copyright in the Dissertation held by the Author.
Microform Edition © ProQuest LLC.
All rights reserved. This work is protected against
unauthorized copying under Title 17, United States Code
uesf
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, Ml 48106- 1346
U N IVER SITY O F S O U TH E R N C A LIFO R N IA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 9 0 0 0 7
fh-P.
Ps
'J?
This dissertation, w ritten by
Felicia Friendly
JlS‘ 4 - ’ gÇ^
M.
under the direction of /i.er,....... Dissertation
Committee, and approved by all its members, has
been presented to and accepted by the Dean of The
Graduate School, in partial fu lfillm ent of the re
quirements fo r the degree of
DO CTO R OF PH ILO SO PH Y
Date
Dean
DISSERTATION COMMITTEE
Chairrmn
ACKNOWLEDGMENTS
I owe a debt of gratitude to Dr. Stephen E. Berger,
Chairman of my committee, who guided and supported me from
the inception of this project to its completion. His
generous help, most valuable advice and above all, his un
dying faith in my ability to succeed, inspired my work at
every stage. I also wish to thank Dr. Joseph Pruitt and
Dr. Betty Walker, my other committee members, for their
help and assistance; and Miss Lynne11 Bennett-Harris who
helped in my data collection. In addition, I wish to thank
the following institutions for their cooperation:
The National Fellowships Fund, Atlanta, Georgia for its
financial assistance; Psychology Department, University of
Southern California; Psychology Department, Los Angeles
City College ; and the Pan African Studies Department,
California State University at Los Angeles.
Finally, I wish to record my debt to the following
persons for their full support and encouragement: My
parents, George and Bernice ; my siblings, Donel, Mickel,
Vanessa, Chesley, Renee, Sabrinia and Edward; and last but
not least, Chris Thomas, for his emotional support.
11
TABLE OF CONTENTS
Page
LIST OF FIGURES .......... iv
LIST OF TABLES . v
Chapter
I INTRODUCTION ................ 1
Conceptual Learning and Age ......... 2
Stimulus Complexity and
Conceptual Learning ...................... 5
Effects of Instruction .... 7
Mediation and Nonmediation................ . . 10
Conceptual Learning Tasks and
Methods of Presentation .......... 15
Intellectual Functioning, Symbolic
Information Processing and Race ...... 21
II RESEARCH DESIGN........... 28
III HYPOTHESES .... ...... 32
IV METHO D............................... 34
Subjects............... 34
Equipment ............. 35
Learning Tasks and IQ Measures ............. 35
Family Background Information ........ 43
Procedure............................... 4 3
Independent Variables . 45
Dependent Variables ........................ 4 6
V, RESULTS.................................... . 47
Analysis of "Soft"....................51
Analysis of "White"............. 57
Comparisons of "White"
and "Soft"...........................59
Analysis of Paired Associate
List Performance...................... 62
111
Chapter Page
VI DISCUSSION................. 66
Evaluation of Hypotheses ........... 68
Mediation and IQ Scores............. .. 72
Mediation and Black Students ............... 74
"White" and Black Students .................. 76
Limitations and Implications for
Future Research ........................... 8 0
REFERENCES............................................ 8 3
APPENDIX................................. ............. 8 7
IV
LIST OF FIGURES
Figure Page
1 GRAPHIC PRESENTATION OF THE INTERACTION OF
IQ SCORE AND CONDITION FOR THE CONCEPT
"SOFT" . 56
2 GRAPHIC PRESENTATION OF THE INTERACTION OF
IQ SCORE AND CONDITION FOR THE CONCEPT
"WHITE"................................. .. 60
3 GRAPHIC PRESENTATION OF THE THREE-WAY
INTERACTION OF IQ SCORE, CONDITION AND
CONCEPT.......................................... 6 3
V
LIST OF TABLES
Table Page
1 PAIRED ASSOCIATE LIST 1 . 37
2 PAIRED ASSOCIATE LIST 2 . . . 38
3 CONCEPT FORMATION LIST - ORDER 1 ...............4 0
4 CONCEPT FORMATION LIST - ORDER 2 .................41
5 MEANS OF THE NUMBER OF TRIALS FOR
ALL IQ GROUPS TO LEARN BOTH CONCEPTS
IN THE MEDIATION AND NONMEDIATION
CONDITIONS . 4 8
6 MEANS OF THE NUMBER OF TRIALS TO
LEARN BOTH CONCEPTS IN THE MEDIATION
AND NONMEDIATION CONDITIONS AVERAGED
ACROSS IQ GROUPS............. 49
7 MEANS OF THE NUMBER TRIALS TO LEARN BOTH
CONCEPTS FOR ALL IQ GROUPS AVERAGED
ACROSS CONDITIONS ......................... . 50
8 MEANS OF THE NUMBER OF TRIALS TO LEARN
CONCEPT "SOFT" IN THE MEDIATION AND
NONMEDIATION CONDITIONS FOR IQ GROUPS ........... 52
9 MEANS OF THE NUMBER OF TRIALS TO LEARN
CONCEPT "SOFT" IN THE MEDIATION AND
NONMEDIATION CONDITIONS FOR ALL IQ
GROUPS ««•53
10 MEANS OF THE NUMBER OF TRIALS TO LEARN
CONCEPT "WHITE" IN THE MEDIATION AND
NONMEDIATION CONDITIONS FOR ALL IQ
GROUPS .......................... 58
11 MEANS OF THE NUMBER OF TRIALS REQUIRED TO
LEARN PAIRED ASSOCIATE LISTS FOR ALL IQ
SCORES ........ 65
VI
CHAPTER
INTRODUCTION
In previous studies of the relationship of IQ test
scores to conceptual learning, researchers have found that
performance is correlated with the interaction of numerous
subject variables as well as the organization and structure
of the learning tasks. Among demonstrated, critical
factors are: 1)the kind of individual difference variable
investigated (e.g., a score on an IQ test or creativity
measure); 2)the particular measuring instrument used (e.g.,
the Wechsler Adult Intelligence Test); 3)the task employed
(e.g., mediation or nonmediation paradigm; 5)the performance
indicator employed (e.g., trials or errors); 6)the informa
tion provided by the task (e.g., amount of relevant or ir
relevant information provided); and 7)the stage of the learn
ing process investigated (e.g., time to the first correct
response or time from the first correct response to concept
attainment (Jacobson, Dickinson, Fleishman and Haraguchi,
1969; Jacobson, Elenewski and Lordahal, 1968; Jacobson,
Elenewski, Lordahal and Liroff, 196 8; Jacobson, Millham and
Berger, 1969; Jensen, 1968; Osier and Fivel, 1961; Osier
and Trautman, 1961; and Osier and Weiss, 1962).
Conceptual Learning and Age
Results of the individual studies appear to be de
termined by the specific interactions of the experimental
modifications in each of these factors. That is, all the
factors just mentioned influence the results of any study.
For example. Osier and Fivel (19 61) investigated the re
lationship of intellectual functioning and age in concept
attainment employing,' the inductive method: Subjects were
presented with different stimuli, all of which belonged to
a common category, and were expected to give the same re
sponse to each element of the group. Prior to this study,
researchers had employed the interview-questionnaire method
or the performance method (Vinacke, 1952).
Osier and Fivel (1961) felt that both of these
methods presented problems since the former assumed that
coexisting with a child's knowledge of a concept was his
ability to translate this knowledge of a concept into words
and the latter was essentially a sorting technique in which
performance could be determined by factors other than knowl
edge of the specific concept under observation. For in
stance, the interest, experience and preference of the sub
ject were important determining factors in concept learning
where alternative responses existed for categorizing. With
the employment of the inductive method, however, correct
attainment of concepts is achieved solely when the subject
responds to the new stimuli at a level of accuracy pre- _
scribed by the experimenter. Consequently, reducing or
eliminating the power of competing stimuli.
Furthermore, the Osier and Fivel (1961) investiga
tion contrasted normal and superior subjects while previous
studies of conceptual learning compared normal and feeble
minded subjects. Since the subjects represented the ex
tremes and midpoint of the elementary school population
and were included within Piaget's three stages of mental
development, the subjects were within three age groups,
6-, 10-, and 14-,years. Three concepts (bird, animal and
living thing) were used because they represented increasing
levels, of difficulty which would increase the discrimina
tive power of the tasks (Welch and Long, 1940). Subjects
were shown two wooden boxes with a picture attached to
each, one depicting the desired concept and the other un
related or neutral to the concept. Instructions for suc^
cessful completion of the task were not explicitly given.
Subjects were merely told that a marble was hidden under
one Of the boxes and if they looked at the picture they
would choose the correct box every time and win the marble.
The Wechsler Intelligence Scale for Children was adminis
tered as the IQ measure.
Results indicated that age and IQ scores were as
sociated with more effective concept learning while no ef
fect of concept difficulty was apparent. However, there
was a nominal effect of concept hierarchy displayed for
subjects who scored in the normal range.
Merely knowing that age and IQ scores were cor
related with concept attainment was not sufficient to clari
fy the specific mechanism by which these two varibles
enhanced performance. Osier and fivel (1961) wanted to
ascertain if there was a qualitative difference in the
learning strategies of older and higher scoring subjects
or simply a greater increase in response tendency. Find
ings of Kendler and Kendler (1959) suggested that to examine
this question one had only to observe the learning curves
of subjects and one would discover a gradually inclining
curve for lower scoring subjects and a curve which increased
suddenly for higher scorers. A gradual attainment of a
concept would be compatible with continuity theory of learn
ing whereas sudden acquistion would be compatible with the
theory of learning through hypothesis testing. In an at
tempt to explore this phenomenon. Osier and Fivel (19 61)
analyzed the responses just prior to concept attainment of
their subjects. These curves, indeed, were demonstrated to
exist for the lower and higher scoring groups but neither
age nor concept hierarchy yielded any difference. There
fore, the correlation between concept attainment and IQ
scores across ages appeared to be the result of hypothesis
testing.
Stimulus Complexity and
Conceptual Learning
Osier and Trautman (1961) further investigated the
Kendler and Kendler (1959) hypothesis that high and low
scorers on an IQ test utilized symbolic mediators in dis
crimination learning and reached concept attainment by
building and testing successive hypotheses, whereas low
scoring subjects were inclined to learn through and build
up S-R associations. Consequently, if one were to observe
the learning curves of high and low scoring subjects, it
was hypothesized that one would discover a gradually in
clining curve for low scoring subjects and a curve which
increased suddenly for high scorers. , Furthermore, it was
contended that if the hypothesis of the varying learning
mechanisms was valid and high scorers employed hypothesis
testing in concept attainment, it should be possible to in
fluence their performance by altering the number of relevant
or irrelevant dimensions on which hypotheses could be
founded; low scorers ’ were hypothesized to be unaffected by
the amount of relevant or irrelevant stimuli since they
were believed to learn through S-R associations.
The investigation was of a single concept: The
number two represented by two sets of materials ; one set
represented the single concept while the other set repre
sented the same concept but was designed to suggest many
more hypotheses than the first set by containing objects
varied in shape, size, color and content. The positive
stimuli of the first set. Formal Two (FT), consisted of
two solid, black circles positioned in random, unduplicated,
patterns on 5 X 5 cards. The negative stimuli of this set
consisted of identical black circles in numbers ranging
from one to five, excluding two, also positioned random
patterns on the cards.
The positive stimuli for the second set contained
two identical pictures of common objects on cards identical
to those used in the FT set. The pictures varied in shape,
size, color and content. This set of stimuli was called
the Object Two (OT) set. The negative stimuli for this
set consisted of the same kind of pictures but were in
numbers other than two. -
The rationale for the design of the stimuli was
that the OT set would suggest many more hypotheses than the
FT set due to the assorted shape, size, color and content
of the various objects depicted. Consequently, the per
formance of higher scoring subjects would be affected since
they were believed to utilize hypothesis testing strategies
in concept attainment. Lower scoring subjects, however,
would not be affected by varying the number of irrelevant
dimensions on which hypotheses could be tested for they
would tend to achieve solution of the problem through the
gradual building up and learning of S-R associations. More
specifically, it was hypothesized that results would show a
6
significant interaction of IQ test scores and the number of
irrelevant dimensions in the concept sets.
Subjects were elementary and junior high school
children. Each subject worked individually on either the
OT or FT set. Subjects were shown the positive and negative
stimuli of a set together and asked to choose the correct
one of the two. The criterion of success was ten consecu
tive correct responses.
Results indicated that subjects who scored high on
the IQ measure showed more effective concept learning than
lower scoring subjects on the FT task where information
which was useful in symbolic information processing was
provided. However, in the similar OT concept learning
situation in which the amount of irrelevant information
was increased, no significant difference between high and
low scoring subjects' performance was found. Therefore, a
significant interaction of test scores and mediation was
found, supporting the hypothesis that high and low IQ scor
ing subjects utilize different learning strategies.
Effects of Instruction
In addition to studies on the amount of relevant
information provided, effects of instructions, mediation,
and IQ scores have been studied as well '(Osier and Weiss,
1962). It had been hypothesized that the superior perfor-
7
mance of higher scoring subjects might be the result of
their supplementing the experimenter's instructions with
their own instructions to themselves, instructions which in
duce them to search for consistencies and test hypotheses.
More specifically, when instructions are vague, higher
scorers would tend to perform better than lower scorers, but
this advantage would not be as apparent as the instructions
became more specific and there would be less need for prob
lem finding. In essence, it was postulated that with non
specific instructions learning tasks presented to subjects
have two components, namely, problem finding and problem
solving; and that the superior performance of higher IQ test
scorers might be attributable to greater proficiency at
either or both of these components. On the other hand,
when instructions are specific and the problem which the
subject is to solve is defined, the problem finding com
ponent was believed to be reduced to a minimum or resolved,
thus, diminishing the superiority of the higher IQ scorers.
Two experiments were performed, both replicating
the methodology of the Osier and Fivel (1961) study. Experi
ment II was identical to Experiment I except for the speci
ficity of the instructions rendered. Subjects in Experi
ment I were told that if they were to look at the pictures
they would be able to get many marbles. On the other hand,
subjects in Experiment II were told that if they looked at
the pictures carefully they would see that there was some
8
thing in the pictures like an idea that would tell them
which one to choose to get a marble every time. The criter
ion for success was 10 consecutive correct responses. At
the completion of the task, subjects were asked how they
knew the correct answer.
Results showed that with instructions which did not
specify the nature of the task, age and IQ score were sig
nificantly correlated with concept attainment, with older,
higher scoring subjects performing better. However, when
the instructions as to the nature of the task were ex
plicitly given, the effect of IQ test score was no longer
apparent and only age was a significant predictor of per
formance. Similar results were obtained when the number of
subjects who were able to verbalize the concepts were
analyzed.
Osier and Weiss (1962) concluded that if their as
sumption that specific instructions reduced or eliminated
the problem finding aspect of the learning tasks, it could
De inferred that:
. . . higher intelligence gave S an advantage
in problem definition but not in problem solu
tion in the concept attainment task.
Another valid possibility, however, is that problem finding
abilities subsequently gave subjects the advantage when
taking the IQ test. Thus, problem finding abilities may be
a prequisite to scoring high on an IQ measure instead of the
9
converse.
Furthermore, it was concluded that if the concept
tasks had been made more difficult, the high IQ scorers
might also have shown superior performance in the problem
solving phase of the task. In addition, since age was
found to be significant in concept attainment, maturation
was postulated to entail growth in problem solving ability
at least until some point between 10 and 14.
To determine the changes in problem finding ability
as it relates to age, performance at each age level under
the two instructional sets were compared. Results suggest
ed that between 6 and 10 years subjects' performances were
enhanced by explicitness of instructions. It appears that
the superior performance attributed to age was not the
result of improvement in problem finding but by more ef
fective problem solving.
Mediation and Nonmediation
In the above studies, the relationships between
intelligence, mediation and nonmediation were not directly
controlled but inferred from observation of subjects' per
formances on a variety of tasks. Several assumptions con
cerning the relationship of intellectual functioning and
the use of conceptual processes were made in these studies:
1)Individuals who score high on an IQ test learn concepts
10
more readily than persons who score lower; 2)High scorers
use mediation and other cognitive processes more effec
tively than lower scorers ; and 3)Intellectual functioning
operates differentially at various stages of the learning
processes. However, these hypotheses had not been
validated empirically (Jacobson, Dickinson, et al., 1969).
Jacobson, Dickinson, et al., (1969) , however, ad
ministered the Cooperative School and College Abilities
Test (SCAT) and employed a paradigm in which mediating
processes could be directly controlled. The paradigm con
structed by Mednick and Freeman (1960) in which the non
mediated paradigm was modified so as to simultaneously
study both mediation and nonmediation was employed.
Subjects (white, undergraduate volunteers) learned
one of two paired associate lists and a concept list. The
paired associate lists were composed of 12 word pairs. On
the first list (PAL-1) 4 of the 12 word pairs were designed
so as to facilitate the subsequent learning of the con
cept "white" on the concept list. The other 8 word pairs
were neutral with regard to "white." The second paired
associate list (PAL-2) was similar to the first list but
4 of its 12 word pairs were designed to facilitate the
later learning of the concept "soft." The concept list
contained 12 nouns which could be placed into three groups
of four words each, with the four words in each group
representing a common concept. Subjects were required to
11
learn the three adjectives ("white," "soft," and buffer
concept "large"), each of which described a different
group of the 4 nouns. For those subjects who learned
PAL-1, the learning of "white" on the concept list was
facilitated, and the concept "soft" was facilitated for
those subjects who learned PAL-2. The concept "large" had
not been facilitated on either of the paired associated
list and was included as a buffer.
Mediation and nonmediation concept scores were ob
tained for each subject; with the mediation score being
derived from the subject's responses to the 4 words which
were related to the facilitated concept on the paired as
sociate list learned. The nonmediation score was derived
from performance on the 4 words related to the concept
which was not facilitated on the paired associate list.
Therefore, each subject was his own control.
Paired associate lists were learned to a criterion
of two successive perfect trials, and the concept list was
learned to a criterion of one perfect trial. The word
pairs and concept list items were presented at 4-second
intervals. Three arrangements of the lists were utilized
and kept constant for all subjects so as to control for
presentation effects. The SCAT was administered at the
end of the session.
The results of the study suggested that 1)utiliza
tion of the mediation paradigm improved the performance of
12
all subjects, supporting the assumption that mediated
generalization facilitates learning of concepts; 2)sub
jects who scored high on the SCAT learned both mediated
and nonmediated concepts more rapidly than lower scoring
subjects. They did not find a significant interaction of
test scores and mediation conditions, which suggested that
when subjects learn the same materials in mediated and
nonmediated concept learning, the magnitude of the increase
in mediated concept learning is the same for higher and
lower scoring subjects. That is, the facilitative effect
of the mediated concept learning over the nonmediated con
cept learning is a constant, additive function rather than
a multiplicative one. Hence, these findings did not sup
port the hypothesis that higher IQ scorers' superior per
formance over lower scorer's performance could be accounted
for by their superior use of cognitive processes, speci
fically symbolic information processing.
Jacobson, Millham, and Berger (1969) examined the
question: Do individuals of varying intelligence use the
same cognitive processes but with differential efficiency,
or do they employ different processes in retrieving infor
mation in mediated conceptualization?
The methodology used was the same as in the afore
mentioned study. Again, undergraduate volunteers were
used. Instead of the SCAT the Vocabulary test of the
Wechsler Adult Intelligence Scale (WAIS), which is the
13
best single predictor of Full Scale IQ test scores, was
administered.
The results of the study demonstrated that 1)con
cepts were learned significantly more rapidly under the
mediation condition, and 2)in contrast to the earlier
study, IQ scores did interact with the mediation condi
tion. Subjects who scored high learned mediated concepts
more rapidly than subjects who scored low, but this effect
was not found in the learning of nonmediated concepts;
there was no significant difference between the performance
on the nonmediated task of those subjects who scored higher
and those who scored lower. It was concluded that lower
scorers employed the same learning strategies for both
tasks since their performance did not improve from the
nonmediated to mediated tasks, while the higher scoring
subjects solved the mediation problem more efficiently by
the use of symbolic processes. Thus, the two measures of
intellectual functioning (WAIS and SCAT) interacted dif
ferentially with the experimental conditions. That is,
subjects scoring higher on an IQ measure do not always
outperform lower scoring subjects. However, IQ scores and
SCAT scores are correlated only about .6 and thus it is
difficult to access the meaning of the Jacobson,
Dickinson, et al. study. One possibility is that the SCAT
is a good predictor of people who try harder on verbal
learning tasks.
14
Conceptual Learning Tasks and
Methods of Presentation
The procedure most often used in conceptual learn
ing studies has been the method of successive presentation,
also called the response method. In this method subjects
are presented with a series of stimuli and required to
name each stimulus as it appears as an example of a par
ticular concept. Using this method researchers have
designated two stages into which conceptual learning may
be divided. The first stage has been termed the response
generation stage of conceptualization. This is the inter
val prior to the first correct response, where subjects
must retrieve or process information in order to generate
the first correct response. The problem is believed to be
a generative one whereby subjects must use mediation maxi
mally to produce the first correct response (Jacobson,
Dickinson, Fleishman, and Haraguchi, 196 9; Jacobson,
Elenewski, Lordahl, and Liroff, 1968; Jacobson, Millham,
and Berger, 19 69).
The second stage has been differentiated as the
period from the first correct response to concept attain
ment. During this recognition stage of conceptualization,
as it has been termed, the subjects must relate or general
ize the correct response to similar instances.
Jacobson, Dickinson, et al., (1969) , in addition
to exploring the learning strategies utilized by subjects
15
in conceptual tasks, investigated the stage at which IQ
scores and mediation interact. It was hypothesized that
during the response generation stage high scorers would
display their greater superiority.
An analysis of the number of triais to the first
correct response (response generation stage) showed sig
nificant effects of IQ scores and mediation, while an
analysis of the number of trials from the first correct
response to the attainment of concepts (recognition stage)
did not show a significant effect of IQ score and media
tion. It was found that the interaction of IQ scores and
mediation occurs almost exclusively in the response gener
ation stage, indicating that mediation may be associated
with information retrieval rather than recognition or re
sponse identification.
The response method has frequently been thought of
as being identical with or a variant of rote learning
(Colemen, 1964). Jacobson, Elenewski, and Lordahl (1968),
in order to assess this hypothesis, compared the perfor
mances of subjects on mediated and nonmediated variants of
the response method with simple performance on paired\\‘
associate (rote) learning tasks. The methodology adopted
was identical to those previously mentioned. Subjects
were white, undergraduates who learned one of two paired
associate lists and a concept list with the concepts
"white" and "soft" being facilitated by having learned one
16
or the other of the two PAL's. Four indices of paired as
sociate learning were utilized: 1)trials to learn all
pairs; 2)number of omission errors ; 3)number of incorrect
responses; and 4)total number of errors (item 4 = item 2 +
item 3).
Six indices of concept learning were employed:
1)trials to the first correct response; 2)trials from the
first correct response to concept attainment ; 3)total
trials to concept attainment (item 3 = item 1 + item 2);
4)number of omission errors; 5)number of incorrect re
sponses ; and 6)total number of errors (item 6 = item 4 +
item 5).
Based upon the results it was concluded that no
significant relation was indicated between paired as
sociate learning and performance on mediated concept
learning- The hypothesis that the nonmediated task is a
form of rote learning was not supported. In addition, a
small but significant relationship was found between the
mediated and nonmediated lists.
The second most widely used alternative to the
response method is the method of simultaneous presentation
where all instances of a concept are presented together as
a set. The subject is expected to identify simultaneously
all the stimuli as examples of the correct concept. While
the response method sometimes has been thought of as a
variant of rote learning, the method of simultaneous pre-
, . . - . , . 17
sentation has been considered as greatly contingent on sym
bolic processes (Bourne, 1966).
In an effort to determine if the low correlation
demonstrated between mediated and nonmediated lists was
confined to the method of successive presentation or was a
general characteristic of concept learning on a variety of
tasks. MiIlham, Jacobson, and Berger (1971) studied the
relationship of IQ test scores and mediation by utilizing
two variants of the method of simultaneous presentation:
massed practice and distributed practice. In the massed
practice variant, the stimuli of each set were presented
consecutively for a total of 20 trials; in the distributed
practice variant each set of stimuli were presented for a
total of 20 trials but the sets were in mixed rather than
consecutive order.
The paired associate and concept lists were the
same as those used in the previous studies. Subjects
were randomly assigned to either the massed or distributed
practice conditions, both of which were constructed to
differentially facilitate the use of symbolic processes.
The massed practice task was designed to facilitate the
employment of symbolic processing and minimize interference
effects; while the distributed practice task was designed
to interrupt continuous symbolic information processing and
to promote interference effects. In addition, two meas
ures of intellectual functioning were employed. The
18
Vocabulary test of the WAIS and the Abstraction Scale of
the Shipley-Hartford intelligence test. It was felt that
vocabulary and abstraction are indicators of associative
and inferential processing, respectively. Vocabulary
tests are thought to measure the ability of the subject to
respond to verbal stimuli with common (conventional) and
equivalent (synonymous) associative responses. It is
necessary that the responses be both common and equivalent
associates since they have to meet norms found in the
dictionary which dictates those associative responses
that are commonly regarded by the language community as
stimulus equivalent.
Accordingly, scores on a vocabulary test have been
postulated and demonstrated to be excellent predictors of
the use of associative processes in concept learning em
ploying the response method (Jacobson, MiIlham, and
Berger, 1969; Jacobson, Elenewski, Lordahl, and Liroff,
1968) .
In contrast, scores on the Shipley-Hartford Ab
straction subtest has been viewed as indicative of in
ferential information processing ability. For example, the
Abstraction test requires subjects to finish the following
line :
"oh ho/ rat tar/mood......... "
19
Inferring the correct rule for the solution of this problem
is necessary for completion. Differing from the Vocabulary
subtest, it is not necessary for the subject to produce a
stimulus equivalent to the components presented. Thus,
inferential information processing is necessary for the
solution to Abstraction test problems (MiIlham, Jacobson,
and Berger, 19 71). It was further hypothesized that as
sociative and inferential processes could be conceptually
and empirically separated.
Generally, results indicated that both masses
practice and mediation facilitated learning. The joint
presence of both conditions did not, though, produce an
additive effect. More detailed analysis, however, in
dicated that high vocabulary scorers learned concepts
quicker under massed practice in both the mediation and
nonmediation conditions, while there were no significant
differences found between the massed and distributed
practice in either the mediation or nonmediation condition
for lower scoring subjects. It was concluded that sub
jects who are high in associative and/or inferential pro
cessing abilities perform better than lower scoring sub
jects when the task is arranged so as to permit subjects
to utilize information provided by the mediation condition
if time is available for the continual, uninterrupted
processing of this information.
20
Intellectual Functioning, Symbolic
Information Processing and Race
A major controversy in the field of intelligence
testing is the issue of performance of black persons on
IQ tests. Some researchers, such as Jensen (19 6 9), have
argued that the commonly observed difference in scores of
blacks and whites does not reflect differential quality of
education or other environmental variables but rather
reflects a basic, physiological, genetic factor. He pos
tulated that blacks display different and inferior learn
ing strategies, patterns, and styles as compared to
whites. In speaking of these differing modes of learning,
Jensen hypothesized two distinctive types of learning
abilities, associative and conceptual; designated as
Level I and Level II, respectively. Level I abilities
would involve learning whereby relatively little transfor
mation of the input stimuli is required, so there is little
or no difference between the forms of the stimuli and the
forms of the responses. Rote learning, serial learning
and paired associate learning were considered indices of
Level I type learning tasks. Level II abilities, however,
would require the subject to actively transform or mani
pulate the stimuli input through self-initiated elabora
tion before overtly responding. Concept learning and
problem solving tasks were given as examples, with IQ
scores being cited as good measuring devices.
21
Representative of the methodology employed and re
sults obtained in Jensen's investigations is a study (Jen
sen and Rowher, 19 69) where 100 low SES black pre-school
children in a private nursery between the ages of 3 and 5
were given serial learning tests, digit span tests and a
paired associate test, with the paired associate test con
taining pictures of familiar objects. The Peabody Picture
Vocabulary Test was administered as the IQ measure. Based
upon the results of this and other studies (Jensen and
Rohwer, 1963; and Jensen, 1968), the following characteris
tics of the levels were postulated.
LEVEL I ABILITIES
Indicative of associative
learning ability.
Necessary but not suffi
cient for Level II abili
ties .
Evenly distributed in all
social class groups (SES)
4. Neural structures necess
ary for Level II abilities
not present.
LEVEL II ABILITIES
1. Indicative of concep
tual and problem
solving abilities.
2. Differentially dis
tributed in the lower
class and middle class,
with lower classes
having less of this
ability.
3. Certain neural struc
tures must be present
for this ability to
develop; these struc
tures have become dif
ferentially distributed
in the population.
4. Most important for
scholastic achievement.
22
As a consequent of the hypothesized differing
nature of intellectual functioning of blacks and whites,
Jensen felt that the process of educating blacks should be
different and directed toward the learning of more speci
fic skills, since their strengths are not believed to be
of a cognitive or "intellectual" type. The educational
achievements of these children could be raised through
instructional methods which would make the attainment of
scholastic skills less dependent upon Level II and en
compass Level I abilities more (Jensen, 1972).
A recent study by Friendly and Berger (19 76) was
designed to examine the question of black-white differen
ces in the nature and quality of intellectual functioning.
Their study provided a direct comparison of the effective
ness on intellectual functioning in black and white first
grade children. In order to develop a viable construct
for intellectual functioning, it was necessary to extend
these findings across the developmental age span. Thus,
another intention of the Friendly and Berger study was to
extend the findings with college students regarding media
tion and intellectual functioning to first graders. That
is, they wanted to see if mediation and IQ scores operated
in the same manner for first graders as reported above for
college students.
The methodology used with college students was
adapted in use with the first graders, providing a continu-
23
ity with the previous studies. Since first graders cannot
read, stimuli consisted of combinations of letters from
the alphabet and pictures of common objects.
Paired associate lists were constructed using a
letter of the alphabet paired with a picture of common
object be effective in enhancing isymbolic information
processing on a variety of tasks.
One of the main purposes of the Friendly and
Berger (1976) was to determine if the interaction of
mediation and IQ scores found for college students could
be replicated using first graders as subjects. In that
study, as in the studies with college students, a vocabu
lary measure of intellectual functioning was used. The
significant interaction of mediation and IQ scores was in
fact replicated with the first graders. Apparently, only
the higher scoring subjects utilized the mediation condi
tion to sufficiently improve their performance. On the
initial paired associates task higher IQ and lower IQ
scoring subjects did not demonstrate differential symbollie
information processing. The results of this study clearly
demonstrated that symbolic information processing operates
similarly across ages in subjects who score higher on IQ
tests, and that verbal mediation plays an important role
in the performance of high IQ scoring subjects.
The results just reported were obtained for both
black and white first graders. That is, for both black and
24
white children scoring higher on the IQ measure, paired
associate learning was best under the mediation condition
and better than for children scoring lower on the IQ
measure. In addition, there was no difference in per
formance on the initial, non-mediated paired associate
task between higher and lower scoring children within
either the black or white samples (an outcome (e.g., letter
A with a picture of a man). All subjects learned this non
mediated list of six pairs to a criterion of two success
ive, perfect trials. On a second paired associate task,
half the subjects received a list on which performance
was mediated from learning the first list (e.g., letter A
with a picture of a woman). The other subjects received a
non-mediated list for their second task (e.g., letter A
with a picture of a block). The Peabody Picture Vocabu
lary Test was used to obtain IQ scores and to obtain equi
valently scoring groups of black and white children.
Consistent with findings on college students, there
was no significant difference in performance between higher
and lower IQ scorers on the first, non-mediated task.
Both black and white higher IQ scorers showed significant
improvement from the non-mediated to the mediation task.
These findings also paralleled those for college students.
Finally, both black and white higher scorers performed
better than lower scorers on the mediated list, again
paralleling college students.
25
It was found that performance was best under medi
ation conditions. This finding is consistent with the
results of Jacobson, Dickinson, Fleishman, and Haraguchi,
1969; and Jacobson, Millham, and Berger, 1969. In the
Friendly and Berger (19 7 6) experiment, performance was
facilitated on paired associate learning task whereas
earlier studies used concept attainment tasks. Thus,
mediation operates for both black and white higher scoring
groups^which contrasts with Jensen's 19 69 hypotheses. He
believed that blacks were high in associative abilities.
Following his reasoning, one would expect black subjects
to perform significantly better than whites under the non-
mediated paired associates task but to perform significant
ly less well under the mediation condition. Such results
simply were not obtained. As with white subjects, high
scoring black subjects utilized the mediation condition
to sufficiently improve performance. Thus, the nature of
intellectual functioning was demonstrated to be the same
for both black and white children. While Jensen (19 69)
accused the proponents of the environmental basis of dif
ferences in IQ scores of blacks and whites of hypothesizing
upon ad hoc data and making generalizations, he, too, made
the same mistake of generalizing from inadequate and faulty
data. The Friendly and Berger (19 76) study clearly indi
cates that the nature of intellectual functioning is the
same for both black and white children.
26
Therefore, despite the claims of some, such as
Jensen (1969) and Shockley (1971), different educational
experiences are not required for some ethnic groups of
children. That is, they do not have to be excluded from
the mainstream of American education. If the nature and
quality of intellectual functioning of both groups are the
same, then it seems logical to conclude that black children
are no less "educable" than white children and less gene
tically capable of abstract thinking and symbolic informa
tion processing. While it is possible that some black
children may enter the classroom less prepared than white,
middle-class and upper-class children due to environmental
and other factors, this is not an indication of their ul
timate intellectual capabilities nor a prerequiste to their
being placed apart in special classrooms.
27
CHAPTER II
RESEARCH DESIGN
The present study attempted to investigate the
relationship of intellectual functioning, information
processing, and mediation in conceptual learning among
black, college students. Subjects learned one of two
paired associate lists and a concept list. On the concept
list were names of common objects and subjects were ex
pected to learn adjectives to describe characteristics
which they have in common (Underwood and Richardson,
1956) . Certain items on the concept list were mediated by
the subject having learned one or the other of the paired
associate lists.
The methodology employed was the same as that
commonly utilized in conceptual learning research. Namely,
the subjects learned one of two paired associate lists and
a concept list whose components were originally designed by
Mednick and Freeman (1960). The first paired associate
list (PAL-1) was designed to facilitate the attainment of
the concept "white" when the subjects later learned the
concept list. Four of the 12 word pairs on PAL-1 were
designed to facilitate the learning of "white" in the sub
sequent concept attainment task. The remaining word pairs
28
were neutral with regard to "white." The second paired
associate list (PAL-2) was similarly designed but was
constructed to facilitate the attainment of the concept
"soft" in subsequent concept attainment. The concept list
was composed of 12 nouns which were placed into three
groups, each containing four nouns which were instances of
a common concept. Four of the words (one group) were in
stances of the concept "white," four were instances of the
concept "soft," and four were instances of the concept
"large" which was not facilitated on either of the two
paired associate lists but which was included as buffer
concept. The subjects' task was to discover the three ad
jectives ("white," "soft," and buffer concept "large")
which described the different groups of 4 nouns.
For those subjects who learned PAL-1, the learning
of "white" on the concept list was facilitated; and for
those subjects who learned PAL-2, the learning of "soft"
was facilitated.
Mediation and non-mediation scores were derived
for each subject. Mediation scores were derived from the
subject's responses on the four words which were facilitat
ed on the previous PAL list which he learned ; and the non
mediation score was derived from the four words which were
not facilitated on the earlier list. Therefore, each sub
ject served as his own control.
29
Bourne (1966) suggested that learning under the
simultaneous presentation method where all instances of a
concept are presented together as a set is greatly contin
gent upon symbolic information processing. In addition,
Jacobson, Millham, and Berger (19 71) demonstrated that the
massed practice variant of the method was most facilita-
tive for information processing. For these reasons, the
present study employed the massed practice variant of the
simultaneous presentation method, i.e., the subject was ex
pected to identify simultaneously all the stimuli as ex
amples of the correct concept, and the stimuli of each set
were presented in consecutive rather than mixed order.
In addition, to control for possible order effects, two
orders of the concept list were employed. That is, either
"white" or "soft" was the first concept presented to the
subject on the concept list.
Black college students were employed as subjects
since all the previous studies, with the exception of the
Friendly and Berger study (1976) , utilized white college
students. Further, in light of findings by researchers
such as Jensen (1969) and Shockley (1971), it was neces
sary to extend these findings to other racial groups.
Employed in the Friendly and Berger (19 76) study were black
and white first graders, but it was necessary to extend
these findings on intellectual functioning, mediation, and
race to black college students to complete the racial and
30
developmental age spectrims.
In addition, studies have demonstrated that race
of the experimenter is an important variable which influ
ences the performances of subjects on a variety of tasks,
especially learning and intellectual type tasks. As a
consequent, black experimenters were employed only.
Both were females ages 22 and 25 years of age.
The vocabulary subtest of the Wechsler Adult
Intelligence Scale was used in earlier works of conceptual
learning and intellectual functioning because research
indicates that it is the single best predictor of full
scale IQ scores (it correlates with full scale IQ test
scores between 0.7 and 0.9 across age levels), and of
performance on mediation tasks. Finally, since this study
was essentially an extension of the Jacobson, Millham,
and Berger (19 66) study to the intellectual functioning of
blacks, it was most appropriate to use the vocabulary sub
test here as well.
31
CHAPTER III
HYPOTHESES
Information processing strategies of subjects
differing in scores on IQ measures demonstrated that sub
jects who scored lower did not perform significantly
different on mediated and non-mediated tasks, which seemed
to suggest that they utilized the same information pro
cessing in learning both concepts; while those subjects
who scored higher performed significantly better on the
mediated concept list, suggesting that symbolic informa
tion processing might have been used by these individuals.
In addition, it has been demonstrated that vocabulary test
scores are excellent predictors of mediated, but not
non-mediated performances (Jacobson, Millham and Berger,
19 69; Jacobson, Elenewski, Lordahl, and Liroff, 19 68;
Friendly and Berger, 1976).
It was expected, therefore, that both higher and
lower scoring subjects would use similar information pro
cessing strategies in non-mediated paired associate learn
ing. That is, there would be no significant difference in
performance between the higher scoring group and the lower
scoring group on the non-mediated items of the concept
list.
32
Hypothesis 1: There would not be a significant
difference in the mean number of trials required
for the highest scoring group and the lowest
scoring group to give the correct response to the
non-mediated items of the concept list.
In mediated concept learning, however, higher
scoring subjects were expected to utilize symbolic infor
mation processing to enchance their performance, while
lower scoring subjects were expected to employ the same
learning strategies as those used in non-mediated learning
(Jacobson, Millham and Berger, 1969). Therefore, higher
scorers were expected to perform better than l o w e r scoring
subjects under the non-mediation condition; and only the
higher scoring subjects would demonstrate a significant
mediation effect.
Hypothesis 2 : There would be a significant dif
ference in the mean number of trials required for
the highest scoring group to give the correct re
sponse to the mediated items of the concept list.
Hypothesis 3: The mean number of trials required
for the highest scoring group to give the correct
response to the mediated items would be signifi
cantly less than the mean number trials required
to give the correct response for the non-mediated
items.
Hypothesis 4 : There would be no significant dif
ference in the mean number of trials required for
the lowest scoring group to give the correct re
sponse to the mediated and non-mediated items.
33
CHAPTER IV
METHOD
Subjects
Subjects were 46 black females and 14 black males
enrolled at the University of Southern California,
California State University at Los Angeles, and Los Angeles
City College, with the majority of the subjects attending
the University of Southern California. Subjects' ages
ranged from 17 to 24 years, with 19 being the average age.
The average educational levels for the mothers of
the high, medium, and low IQ scoring groups were 15.3, 13.7,
and 12.8, respectively; and for their fathers, 13.9, 13.6,
and 12.4, respectively. The average number of children in
the family were 3.5, 3u,2, and 4.7 and the reported average
incomes were $27,000, $23,000 and $11,000, with more of the
incomes of the lowest group being contributed to by only
one parent. That is, many more of the lowest scoring
group's families than the highest and middle scoring
groups' families were either single parent families or one
of the two parents was unemployed. Too, more of the
highest scoring group's parents were professionally em
ployed, with the most frequently occurring occupation being
34
that of teaching; while the lowest scoring group contained
skilled and unskilled positions with unemployment being
common. The middle scoring group was a combination of
the other two groups.
The average Grade-Point-Averages for the high,
medium, and low groups were 2.9, 3.0 and 2.9 respectively.
Basically, there were no differences in the college majors
of the three groups, with pre-professional majors occurring
most frequently, i.e., pre-law, pre-dentistry, pre-medical,
and pre-pharmacy, etc. However, among the people not
majoring in the aforementioned areas, more of the lowest
scoring subjects were involved in nursing, social work,
and psychology; while the highest scoring subjects were
involved in business administration and journalism.
Equipment
A slide projector with an automatic timer was used
for displaying the slides.
Learning Tasks and IQ Measures
All subjects learned one of two paired associate
lists and one of the two orders of a concept list. Both
paired associate lists consisted of 12 word pairs. On the
first paired associate list (PAL-1), the learning of 4 of
the 12 word pairs, SUGAR-BLACK, HAILSTONE-SHEEP, BUTTON-
35
BREAD, and CABBAGE-SALT, was designed to facilitate the
attainment of the concept "white" in subsequent concept
learning. The remaining 8 word pairs on this list were
neutral in regard to "white." The list is provided in
Table 1.
The second paired associate list (PAL-2) contained
4 word pairs, PUP-HARD, GRAPE-LOUD, PEACH-SMOOTH, and
CRADLE-COMFORT, which were designed to facilitate the at
tainment of the concept "soft" on the subsequent concept
list. The list is provided in Table 2.
The concept list contained 12 words which were
to be placed into 3 groups of 4 words each. The elements
in each group were instances of the same concept. The
subject was required to learn the three adjectives, each
describing one of the three groups. The nouns SUGAR,
HAILSTONE, BUTTON, and CABBAGE were instances of the con
cept "white"; PUP, GRAPE, PEACH, and CRADLE were instances
of the concept "soft"; and ANCHOR, CAMEL, FOREST, and
BARREL were instances of the concept "large." The latter
concept was not facilitated on either of the paired as
sociate lists and was included as a buffer concept. In
addition, two orders of the concept list were employed.
Half of the subjects were presented with "white" as the
first concept to be learned and "soft" as the second;
while half the subjects were presented with "soft" as the
36
TABLE 1
PAIRED ASSOCIATE LIST I
Pair Relevant Assoc.
sugar-black black-white
(.751)^
hailstone-sheep sheep-white (.009)
button-bread bread-white (.015)
cabbage-salt salt-white (.006)
pup-afraid none
grape-swift none
peach-beautiful none
cradle-anger none
anchor-king none
camel-dream none
forest-command none
barrel-lion none
^Probability of occurrence of response term (e.g.,
"white") to stimulus term (e.g.. "black").
37
TABLE 2
PAIRED ASSOCIATE LIST 27
Pair Relevant Assoc.
sugar-afraid none
hailstone-anger none
button-beautiful none
cabbage-swift none
pup-hard hard-soft (.674)^
grape-loud loud-soft (.541)
peach-smooth smooth-soft (.206)
cradle-comfort comfort-soft (.069)
anchor-king none
camel-dream none
forest-command none
barrel-lion none
^Probability of occurrence of response term (e.g.,
"white") to stimulus term (e.g.. "black").
38
first concept to be learned and "white" as the second.
The buffer concept "large" was: the last concept to be
learned for all subjects. The two orders of the concept
list are provided in Table 3 and Table 4.
The process by which the mediation paradigm was
hypothesized to facilitate the use of mediated generaliza
tion by subjects has been explained succinctly by several
researchers (Jacobson, Dickinson, et al., 1969; Mednick and
Freeman, 1960; and Russel and Jenkins, 1954) using the
learning of the concept "white" as an example. In PAL-1
was the word pair SUGAR-BLACK. Mediation theory proposes
that the stimulus word BLACK would evoke the implicit re
sponse "white," since "white" is a common response to
BLACK. This has been schematized by Jacobson, Dickinson,
et al., in the following way:
stimulus presented : sugar-black
implicit response of Subject: white
The implicit response "white" to the explicit word
BLACK causes the word "white" to be present itself in con
tiguity with the word SUGAR in the mind of the subject,
strengthening the association "sugar-white." When the
subject later learns the concept list, the concept "white"
is assumed to be attained more readily than if he had not
been previously presented with the SUGAR-BLACK work pair,
39
TABLE 3
CONCEPT FORMATION LIST - ORDER I
Instance Concept
sugar
hailstone
button "white"
cabbage
pup
grape
peach "soft"
cradle
anchor
camel
forest
"large"
barrel
40
TABLE 4
CONCEPT FORMATION LIST - ORDER 2
Instance Concept
pup
grape
peach "soft"
cradle
sugar
hailstone
button "white"
cabbage
anchor
camel
forest "large"
barrel
41
That is, mediated generalization intensifies the sugar-
white association and thus raises the response dominance
of "white" to the word "sugar."
At the completion of the concept task, all subjects
were administered the Vocabulary subtest of the Wechsler
Adult Intelligence Scale as the IQ test measure since this
test has been found to be the best single predictor of
Full Scale IQ test scores. In addition, it was thought
that the vocabulary test would measure the ability of the
subject to respond to verbal stimuli with what is general
ly considered common (conventional) and equivalent
(synonymous) associative responses, since responses have
to meet specified norms found in the dictionary for maximum
scoring (Jacobson, Millham and Berger, 1971). It was
predicted that the more similar subjects' knowledge of
these conventional definitions the more similarities there
would be in their conceptual learning strategies.
Subjects' vocabulary scores were ranked from high
est to lowest and the 20 subjects with the highest scores
were assigned to the high group, the 20 middle scorers to
the medium group, and the 20 lowest scorers to the low
group. The distribution of verbal IQ scores ranged from
79 to 138, with the high scorers averaging 120, the middle
scorers 10 3 and the low scorers 97.
42
Family Background Information
At the conclusion of this study subjects were asked
to fill out a Family Background Questionnaire. Questions
included their grade level in college, major. Grade-Point-
Average , the number of children in family, mother and
father's last grade completed, mother and father's occupa
tions and the family income. This questionnaire is pro-;l
vided in the Appendix.
Procedure
Each subject was presented with the following
written paired associate list instructions and was told to
read them silently as the experimenter read them aloud.
1. I will present some words paired with some other
words on the screen one at a time. I want you
to look at the pair--look at the first word and
look at the second word— and remember that words
went together as a pair. Because after all pairs
have been presented once, I will show you just
the first word and would like you to tell me what
second word went with it.
2. ZEach first word will be presented for 6 seconds and
you must give a response within the 6 second
period. This will be followed with the correct
pairing (again for 6 seconds) before the next
first word is shown.
3. We will go through the’list and get all word pairs
completely correct twice in a row.
Subjects were then presented with Pal-1 or Pal-2
and required to learn the list to a criterion of two
successive perfect trials. The stimuli were presented at
6 second* intervals on a screen by a slide projector with
an automatic timer.
After reaching the paired associate criterion of
two successive perfect trials, subjects were presented
with the written instructions for the concept list. Again,
they were instructed to read silently as the experimenter
read aloud:
1. There will be 3 groups of 4 nouns presented to you
on the screen; the nouns in each group can be
described by the same word, the second group of
nouns can be described by a different word, and the
last group by yet a different word. Your task
will be to discover the three words which describe
each of the three groups.
2. Each group will be presented for 6 seconds and you
must give a response within the 6 second period.
I will tell you "right" or "wrong" after each
response.
44
3. We will do this for each list several times and
you are to give a response each time the list ap
pears even though you know the response is wrong.
4. When you get the word correct that describes the
4 words in the list, we will go on to the second
group, and when you get that list correct, we will
do the third group.
Subjects were required to learn the concept list to one
perfect trial or terminated after 20 trials. The pro
cedure for display was the same as for the paired asso
ciate list. At the completion of the concept list, each
subject was administered the Vocabulary subtest of the
Wechsler Adult Intelligence Scale, after which they were
asked to complete the Family Background Questionnaire.
Independent Variables
The two independent variables employed in this
study were IQ test score and Condition.
IQ Score was a three level factor : high, medium,
and low, derived from the range distribution in which the
subjects' scores were located.
Condition was a two level factor : mediation and
non-mediation, derived from the subjects' responses to the
items comprising the facilitated and the non-facilitated
concepts on the list he learned, respectively.
45
Dependent Variables
The one dependent variable was Total Trials to
Criterion with two levels:
1. Total number of trials required for subjects to
obtain two successive perfect trials on the paired
associate lists.
2. Total trials required for subjects to obtain the
concepts.
46
CHAPTER V
RESULTS
The data were initially analyzed with a Repeated
Measures Multivariate Analysis of Variance Computer Pro
gram (Clyde, 1969) with Total Trials to Criterion as the
Dependent Variable. An Analysis of Variance was computed
for the Independent Variables of IQ Scores (high. Medium,
low), Conditions (mediation, nonmediation), and Concept
(white or soft).
Table 5 presents the means of the number of
trials required for the higher and lower scoring groups to
attain the concepts in the mediation and nonmediation
conditions for both the concepts "white" and "soft."
Since the concept "large" was included solely as a buffer
concept, it was omitted from the analyses. Observation of
Table 5 indicates that there was a significant interaction
of condition with Concept for Trials to Criterion,
F(l, 12) =: 12. 659, p < .004. These means are averaged
across IQ scores in Table 6 and across condition in Table
7.
Thus, performance under mediation and nonmediation
conditions was dependent upon the Concept under investiga
tion for all subjects regardless of IQ group membership:
47
TABLE 5
MEANS OF THE NUMBER OF TRIALS FOR ALL IQ GROUPS TO
LEARN BOTH CONCEPTS IN THE MEDIATION AND
NONMEDIATION CONDITIONS
CON C E P T
w
Pi WHITE SOFT
tD
CO
Non- Non-
<
CONDITION Mediation Mediation Mediation Mediation
M
S
O
High 15.9 12.8 12. 2 16.9
M
S
;
Medium 9.6 14.0 12.7 14. 2
<
CJ
§
Low 15.0 10. 2 12.8 15.9
48
TABLE 6
MEANS OF THE NUMBER OF TRIALS TO
LEARN BOTH CONCEPTS IN THE
MEDIATION AND NONMEDIATION
CONDITIONS AVERAGED
ACROSS IQ GROUPS
CONCEPT
Condition WHITE SOFT
Mediation 15.4 12.8
Nonmediation 11.6 16.4
49
TABLE 7
MEANS OF THE NUMBER TRIALS TO LEARN BOTH
CONCEPTS FOR ALL IQ GROUPS AVERAGED
ACROSS CONDITIONS
CONCEPT
White Soft
High 15.9 12.2
Low 15.0 12.8
50
Subjects reacted differently to the concepts "white" and
"soft." .Table 5 also indicates that there was a signifi
cant triple interaction among Condition, IQ Scores and
Concept, F(l, 24) = 3.818, p < .06. Consequently,
separate analyses of the two concepts "white" and "soft"
were performed for the high and low IQ groups which were
of interest in this study. Results for the middle group,
however, will be reported when they are contrary to the
others.
Analysis of "Soft"
Table 8 presents the mean number of trials required
for subjects to attain the concept "soft" under the media
tion and nonmediation conditions. It demonstrates that
there was a significant main effect for condition,
F(1, 12) = 12.033, p < .005. That is, subjects required
significantly fewer trials to give the correct response
to the mediated items of the concept list. There was not
a significant effect of IQ scores, F ( 1. 12) = 699, p < .4jL9;
nor was there a significant interaction of IQ and condi
tion, F(1, 12) = 2.242, p < .147. Table 9 presents these
means.
Myers (1975) and McNemar (1969) suggest that the
Scheffe Method of investigating selected contrasts is
appropriate when the overall tests of F demonstrates
51
TABLE 8
MEANS OF THE NUMBER OF TRIALS TO LEARN CONCEPT
"SOFT" IN THE MEDIATION AND NONMEDIATION
CONDITIONS FOR IQ GROUPS
CONCEPT Soft
Mediation 12.5
Non
mediation 16.4
g
I
52
TABLE 9
MEANS OF THE NUMBER OF TRIALS TO LEARN
CONCEPT "SOFT" IN THE MEDIATION AND
NONMEDIATION CONDITIONS FOR ALL IQ
GROUPS
SOFT
CONDITION Mediation Nonmediation
High 12.2 16.9
3
1
Low 12.8 15.9
1
53
significant mean differences and the data suggest specific
comparisons. This ; post hoc method tests specific contrasts
among the means by computing an overall confidence inter
val, K, while keeping the alpha level constant and testing
the mean differences against this Critical Value, K.
Consequently, differences are only regarded as signifi
cant at this prescribed alpha level if they are equal to
or greater than the calculated Critical Value (McNemar,
1969).
Since the data of the present study suggested such
multiple comparisons, Scheffe analyses were performed.
Calculations indicated that 4.32 was the Critical Value
which the mean differences would have to equal or surpass
when the difference was in the predicted direction in
order to be significant, F(l, 60) = 4.00, p < .05; while
6 was calculated to be the Critical Value which the mean
difference would have to equal or surpass to be signifi
cant when this difference was in the opposite direction of
that predicted. Since the alpha level remains constant,
results will be reported as Significant or Not Significant
depending upon whether the Critical Value is equalled or
surpassed.
These analyses indicate :
1. There was not a significant difference in the
mean number of trials required for the highest
54
and lowest scoring IQ groups to give the correct
response, "soft," when "soft" was the mediated
concept.
2. Nor was there a significant difference in the mean
number of trials required for the highest and low
est scoring subjects to give the correct response,
"soft," under the nonmediation condition.
3. When "soft" was the concept either mediated or
nonmediated, the highest scoring group's mean
number of trials required to give the correct
response to the mediated items was (Significantly
less than the mean number of trials they required
to give the correct response to the nonmediated
items of the list. Thus, for the highest scoring
group mediation appeared to exist.
4. While there was a tendency for the lowest scoring
group to benefit from the mediation condition, the
difference between the mean number of trials re
quired to attain the mediated concept and the mean
number of trials required to attain the non
mediated concept did not meet the specified cri
terion for opposite direction significance.
These phenonmena are represented graphically in Figure 1.
55
FIGURE
GRAPHIC PRESENTATION OF THE INTERACTION OF
IQ SCORE AND CONDITION FOR THE
CONCEPT "SOFT
17
16
2 15.
H
p 14-
high
low
11'
Mediation Nonmediation
CONDITION
56
Analysis of "White"
Table 10 presents the number of trials required for
subjects to attain the concept "white" under mediation and
nonmediation. Analyses of these means indicate: ;
1. There was not a significant difference in the
mean number of trials required for the highest and
lowest scoring groups to give the correct response
to the mediated items of the concept list.
2. Nor was there a significant difference in the mean
number of trials required for the highest and
lowest scoring groups to give the correct response
to the nonmediated items of the concept list.
3. Interestingly, for the highest scoring group the
mean number of trials required to give the correct
response to the mediated items of the list was
greater than the mean number of trials required to
give the correct response to the nonmediated items.
Due to the difference in the direction of this
finding and the predicted direction, this differ
ence does not meet the Critical Value required for
significance. That is, when the paired associate
list was constructed so as to facilitate the at
tainment of the concept "white" it took these
highest scoring subjects longer to attain this
concept on the subsequent concept list.
57
TABLE 10
MEANS OF THE NUMBER OF TRIALS TO LEARN
CONCEPT "WHITE" IN THE MEDIATION AND
NONMEDIATION CONDITIONS FOR ALL IQ
GROUPS
WHITE
CONDITION MEDIATION NONMEDIATION
High 15.9 12.8
1
Medium 9.7 14.0
3
g
g
g
Low 15.0 10. 2
58
4. Similar to the performance of the highest scoring
group under mediation, the lowest scoring group
also required a greater number of trials to attain
the mediated concept than the nonmediated concept.
This difference, however, does not meet the criter
ion for significance because of its direction.
These phenomena are represented in Figure 2.
5. Interestingly, the medium scorers required sig
nificantly less trials to attain the mediated con
cepts than either the highest or lowest groups.
6. The medium group did not, however, require
significantly less trials to attain the nonmedia
ted concepts than the highest and lowest scoring
groups.
7. Consequently, unlike the highest and lowest
scoring groups, the medium scorers required sig
nificantly less trials to attain the concept under
the mediation than under the nonmediation con
ditions .
Comparisons of
"White" and "Soft"
As indicated earlier, an Analysis of Variance
demonstrated a triple interaction of IQ score. Condition,
and Concept, F(1, 24) = 3.818, p < .06. Utilizing Scheffe,
Table 5 indicates: gg
FIGURE 2
GRAPHIC PRESENTATION OF THE INTERACTION OF IQ SCORE
AND CONDITION FOR THE CONCEPT "WHITE"
§
s
M
84
M
P : :
o
o
E H
10
M
E H
17
16
15
14
high'
low
1-3
12
11
Nonmediation Mediation
60
The mean number of trials required for the highest
scoring group to reach concept attainment under
mediation when "white" was the concept was greater
than the mean number of trials required for their
highest scoring counterparts who received "soft"
as the mediated concept. This difference, however,
did not reach the Critical Value level.
While approaching significance there was not a
significant difference between the mean number of
trials required for the highest scoring group to
reach concept attainment when "white" was the non
mediated concept and the mean number of trials re
quired to reach attainment when "soft" was the non
mediated concept.
The mean number of trials required for the lowest
scoring group to reach concept attainment under
mediation when "white" was the concept investi
gated was greater than the mean number of trials
required for the lowest scorers who received
"soft" as the mediated concept, but this difference
was not significant.
The mean number of trials required for the lowest
scoring group to reach concept attainment when
"white" was the nonmediated concept was greater
than the mean number of trials required for the
lowest scoring group to reach concept attainment
61
when "soft" was its mediated concept. This dif
ference did not meet the Critical Value for sig
nificance .
5. The mean number of trials required for the medium
group to reach concept attainment under the media
ted condition when "white" was the investigated
concept was not significantly different from the
mean number of trials required to reach concept
attainment when "soft" was involved.
Figure 3 graphically represents these phenomena.
Analysis of Paired Associate
List Performance
Initially, it was believed that an interaction of
which concept was mediated and which concept was presented
first in the concept list might exist; and that the most
conducive situation for mediation to occur would be the one
where the mediated concept, either "white" or "soft," was
also the first concept presented in the concept list, e.g.,
"white" mediated and "white" presented first; or "soft"
mediated and "soft" presented first. This interaction,
however, was not found to be significant.
An analysis of the number of trials required for
the subjects to reach the criterion of two successive
perfect trials on the paired associate list did not indi-
62
FIGURE
GRAPHIC PRESENTATION OF THE THREE-WAY INTERACTION
OF IQ SCORE, CONDITION AND CONCEPT
17
^ high/sOFT
16 low/SOFT
M
H
15
14 U]
H
13
high/WHITE 12
11
low/WHITE
Nonmediation Mediation
CONDITION
63
cate a significant difference in the performance of the
high and low scoring groups, F(l. 38)
These means are presented in Table 11.
64
TABLE 11
MEANS OF THE NUMBER OF TRIALS REQUIRED TO
LEARN PAIRED ASSOCIATE LISTS FOR
ALL IQ GROUPS
High Medium Low
4. 20 4.15 4. 35
65
CHAPTER VI
DISCUSSION
One of the main purposes of this investigation
was to examine the effects of mediation and IQ scores on
conceptual learning for black college students. We at
tempted to determine if the interaction of mediation and
IQ scores found for white college students (Jacobson,
MiIlham, and Berger, 19 76) and black and white first
grade children (Friendly and Berger, 1976) could be re
plicated with black college students. More specifically,
it was predicted that higher IQ scoring subjects would not
outperform lower IQ scoring subjects under nonmediationed
learning conditions but would significantly outperform
lower IQ scoring subjects under mediated learning condi
tions .
In the present study, as in the studies with white
college students, subjects learned one of two paired as
sociate lists and a concept list whose original components
were designed by Mednick and Freeman (1960). The first
paired associate list was designed according to mediated
generalization theory to facilitate the attainment of the
concept "white" on the subsequent concept list. The
second paired associate list was similarly designed but
was constructed to facilitate the subsequent attainment
66
of the concept "soft." The concept list contained three
groups of four nouns each, with the four words in each
group representing instances of the same concept. The
subject's task was to discover the three adjectives which
could describe the three groups. Mediation and non
mediation scores were derived for each subject based upon
his performance on the four words that were facilitated
(mediated) or not facilitated (nonmediated) on the paired
associate list which he had previously learned. At the
conclusion of the concept list the Vocabularly test of the
Wechsler Adult Intelligence Test was administered as the
intellectual functioning measure, followed by a Family
Background Questionnaire to ascertain the socioeconomic
status of the participants.
Results indicated that the two concepts employed
("white" and "soft") interacted differently with the In
dependent Variable, Condition (p < .004). Therefore, it
was decided to perform separate Scheffe Multiple Compari
sons on the two concepts in question. Analyses derived
from the concept "soft" indicated that there was not a
significant IQ score effect but there was a significant
condition effect (p < .005). The condition effect, how
ever, was limited to the mediation condition for the high
est IQ scoring group. Only the highest scoring subjects
performed significantly better under mediation than non
mediation, but they did not significantly outperform lower
67
scoring subjects on the nonmediated task. Analyses
derived from "white" did not indicate a significant main
effect of either IQ score or condition. In fact, there
was reversal of the phenomena found for "soft." That is,
with the concept "white" both the highest and lowest
scoring subjects performed much better under the non
mediation condition and less well under mediation.
Evaluation of Hypotheses
Hypothesis 1. There would not be a significant
difference in the mean number of trials required for the
highest scoring group and the lowest scoring group to
give the correct response to the nonmediated items of the
concept list.
Subjects who scored highest on the vocabulary test
did not significantly outperform the lowest scoring sub
jects on the nonmediated items for either concepts "soft"
and "white." Both groups, though, required less trials to
obtain the nonmediated concept "white" than their counter
parts who received "soft" as the nonmediated concept. In
other words, when "white" was nonmediated it was easier to
learn than "soft" when it was nonmediated for highest and
lowest scoring subjects; while middle scorers learned both
concepts equally under nonmediation.
Thus, for nonmediation, the hypothesis was con-
68
firmed and is consistent with results of MiIlham, Jacob
son and Berger (1971), Jacobson, Millham and Berger (1969),
and Friendly and Berger, 1976).
Hypothesis 2. There would be a significant dif
ference in the mean number of trials required for the
highest scoring group to give the correct response to the
mediated items of the concept list.
Subjects who scored highest on the vocabulary test
did not perform significantly better than the lowest
scoring subjects under the mediation condition for either
concepts "soft" and "white." Thus, for this population of
highest and lowest IQ scoring, black, college students,
the hypothesis that highest scorers would do significantly
better than lowest scorers under mediation was not con
firmed. This finding is also inconsistent with the find
ings of Jacobson, Millham and Berger (1969) but is consis
tent with the findings of Millham, Jacobson and Berger
(1971). These studies will be discussed in greater detail
at a later point in this paper.
In addition, the highest and lowest scoring groups
took much longer to attain the concept under mediation
when "white" was the mediated concept than they took to
attain the concept "soft" when it was mediated. This
finding is in direct opposition tb the finding for nonmedia
tion in the present study. That is, "white" is acquired
69
quicker when it is nonmediated and "soft" is acquired
quicker when it is mediated. This suggests that for these
highest and lowest IQ scoring, black, college students, the
concept "white" is attained if there is not prior experi
ence with it during the learning of the paired associate
list. Under this condition ("white" nonmediated) attain
ment was maximized for this subject population. But per
formance was actually inhibited by exposure to a list
which was designed to facilitate the attainment of "white"
in the subsequent concept list. Middle scorers did not
seem affected by this phenomenon.
Hypothesis 3. The mean number of trials required
for the highest scoring group to give the correct response
to the mediated items would be significantly less than the
mean number of trials required to give the correct response
to the nonmediated items of the concept list.
Highest IQ scoring subjects did, indeed, perform
significantly better under mediation than under nonmedia
tion when "soft" was the investigated concept (p < .05).
Prior experience with a paired associate list designed to
facilitate the attainment of "soft" on the subsequent con
cept list was effective in enhancing the performance of the
highest IQ scoring groups. There was a difference in their
performance under mediation and nonmediation when "white"
was the involved concept, but this difference was in the
70
opposite direction of that predicted. That is, they per
formed better under nonmediation than under mediation.
This finding that highest IQ scoring, black,
college students utilize mediation conditions to enhance
their performance confirms the hypothesis and is consistent
with previous findings of white, college students, and
black and white first grade children. This result com
pletes the developmental age span and suggests that media
tion operates similarly for highest IQ scoring college
students and first graders both black and white. This
is clearly inconsistent with Jensen's (1969) proposals.
Hypothesis 4. There would be no significant dif
ference in the mean number of trials required for the
lowest scoring group to give the correct response to the
mediated and nonmediated items of the concept list.
Similar to the highest IQ scoring group of the
present study, the lowest IQ scoring group did display a
tendency to improve their performance and benefit from
mediation conditions when "soft" was the mediated concept,
which is consistent with the findings of Millham’ , Jacobson
and Berger (19 71) where mediation enhanced performance re
gardless of IQ membership. Unlike the Millham, Jacobson
and Berger (19 71) study, this difference was not signifi
cant. Therefore, while performance improved under media
tion it was not significantly different from their perform
ance under nonmediation. This finding supports 71
the hypothesis.
Mediation and IQ Scores
The finding of a difference, but not a significant
difference, between the performance of highest and lowest
IQ scoring subjects under mediation does not confirm the
hypothesis about mediation effects and how they operate.
Nor is this finding consistent with the findings of Jacob
son, Millham and Berger (1969) and Friendly and Berger
(1976) where significant mean differences between the
groups were demonstrated under mediation conditions. The
lack of a significant difference could be attributed to
the tendency of the lowest IQ scoring group to also profit
from the mediation condition. While the improvement of
this group under mediation was significant, it was suffi
cient to decrease the mean difference between its mediated
mean score and the mediated mean score of the highest
scoring group such that no significant difference between
the two means were found.
Comparison of the primary studies in this area
demonstrates :
1. The Jacobson, Dickinson, et al., (1969) study
indicated significant main effects for IQ and
Mediation but no significant interaction.
72
2. The Jacobson, Millham and Berger (19 69) investiga
tion indicated no significant main effects for IQ
score or Mediation but a significant interaction.
3. The Millham, Jacobson and Berger (1971) study
demonstrated a significant main effect for Media
tion but no significant effect IQ or an interac
tion of IQ score and Mediation.
The results of Millham, Jacobson and Berger
(19 71) seem most comparable to the present findings.
They found improvement in the performance of the lower
scoring population under mediation. It was suggested that
the difference in their results and the results of
Jacobson (1969) was due to the difference in the presenta
tion methods employed in the two studies. The former
utilized the massed practice variant of the simultaneous
method and the latter the successive presentation method.
The present study employed the massed practice variant of
the simultaneous method. Consequently, it is not sur
prising that as in the Millham (1971) study, the lowest
scoring subjects also benefited from mediation. This
suggests that concepts are learned most efficiently.
. . . when the task is arranged so as to provide
simultaneously both the information given by the
mediation condition and time for the continual
uninterrupted processing of the information provi
ded . . . (Millham, Jacobson and Berger, 1971)
73
Mediation and Black Students
While only the highest IQ scoring group utilized
mediation to significantly improve performance, there was
a tendency for the lower scoring IQ groups to profit from
mediation as well. It is possible that the massed prac
tice variant of the simultaneous presentation method may
have operated to enhance the performance of the lower / IQ
scoring black students as well as the performance of the
highest IQ scoring students. Interestingly, when "soft"
was the investigated concept, the mediation and nonmedia
tion scores of the lower IQ scoring black group were less
than the mediation and nonmediation scores of their lower
IQ scoring white counterparts in the Millham (19 71) study.
Review of the personal data collected on the lowest
IQ scoring students may lend clarity to their comparative
ly superior performance. It indicates that IQ group
classification was consistent with socioeconomic status
variables (SES), i.e., parents' education and occupations,
family income and number of children in family. Grade-
Point-Averages, however, appeared independent of IQ group
membership. That is, there was no difference in the school
performance of the three groups. While subjects from lower
SES families knew less words than subjects from higher SES
families, they were able equally to compete academically in
college and university settings. Their performance in the
74
present study would indicate that here, too, they were able
to compete comparably on a variety of intellectual tasks,
i.e., the learning of a paired associate list, and mediated
and nonmediation concepts. It is possible that the lower
scoring black students try harder at achieving intellec
tually than lowering IQ scoring white students. An alter
nate hypothesis might be that vocabulary scores are not
good predictors of Full Scale IQ scores for black college
students as reported for white college students. Since
the vocabulary IQ scores are not predictive of academic
performance for black subjects such scores may not be
good measures of intellectual functioning (ability) for
blacks.
In summation, the findings of the present study
indicate that when "soft" was the concept investigated,
both the highest and lowest scoring groups profited from
the mediation condition; but only the highest scoring
group utilized the mediation condition to sufficiently and
significantly improve performance over nonmediation. This
finding is consistent with previous research on intellec
tual functioning and mediation for white college students
(Jacobson, Millham and Berger, 1969) and black and white
first graders (Friendly and Berger, 1976). This clearly
suggests that the nature of intellectual functioning is the
same for black and white college students as well as for
black and white children. That is, black college students
75
do not display different and inferior learning strategies,
patterns and styles than do their equivalent IQ scoring
white counterparts. Specifically, high IQ scoring black
first graders and college students were shown to effective
ly utilize symbolic information processing.
"White" and Black Students
Findings suggest when highest and lowest scoring
subjects had prior exposure to a list designed to facili
tate the learning of "white" on the subsequent concept
list, there was a resistance to producing the response
"white." This effect was not apparent for "white" when it
had not been mediated on the prior list, i.e., under the
nonmediation condition these subjects were able to produce
the concept "white" much sooner than they had under the
mediation condition.
Minard (1965) suggests when subjects are presented
with emotionally arousing stimuli, a failure to recognize
the stimuli may occur--that differences in the affective
qualities of stimuli affect one's performance. This
phenomenon has been labelled "Perceptual Defense" (John
son, Thompson and Frincke, 1960; McGinnis, 1945; Brown,
1961). McGinnis (1945) succinctly explained perceptual
selection or defense as:
76
. . representing conditioned avoidance of ver
bal symbols having unpleasant meanings to the ob
server. The stimulus word serves as a cue to a
deeply embeded anxiety .... Avoidance of fur
ther anxiety is contemporaneously aroused in the
form of perceptual defense against recognition of
the stimulus object ....
In light of this hypothesis, it is feasible that
the prior experience of black subjects with the concept
"white" during the learning of its mediated paired as
sociate list may have acted as a cue or affective threat
which produced anxiety and consequent cognitive blocking
or perceptual defense. The nature of the black-white
relationship in America, past and present, is such that
one need not search far for an explanation of black stu
dents' possible emotional reactions to the stimulus
"white." In addition to the implicit facilitating of
"white" on its mediated paired associate list, the list
contained the explicit word "black," which may have in
creased the likelihood of a defensive reaction to the
emotion aroused by the concept "white." It would appear,
however, that middle scorers are not as defensive around
this issue or their defenses do not manifest themselves
in this manner.
It should be noted that black students may have
been aware, consciously or unconsciously, of the nature of
the investigation. School regulations regarding human sub
ject participation demanded that an eligibility list (i.e.,
in this case, the names of black students only) be
posted and a general description of the task be included,
i.e., "Learning of Word Pairs Study." While the phrase
"black college student study" was omitted from the formal
description, many of the students were aware that the list
contained only the names of black students and deduced
that it was a black, intellectual functioning study. Con
sequently, students who signed up for the experiment may
have already been made anxious, and the inclusion of
"white" as a mediated concept may have triggered a defen
sive response.
In a recent study. Swinger (1976) investigated the
relationship of militancy, motive to avoid success, and
intellectual functioning. Black women were rated on mili
tancy; half of them were then given the instructions that
"black women do best" and half of them the instructions
"white" women do best" and subsequently administered the
Digit Symbol test and the Shipley Hartford Abstraction and
Vocabulary tests. Results demonstrated a significant main
effect of instructions on vocabulary scores. The black
women of that study performed significantly lower when
told white women performed the task best. In light of
that finding, it is possible that the highest and lowest
IQ scoring subjects of the present study may have been
demonstrating the "motive to avoid success" on the emo
tionally arousing concept, "white."
78
A common argument against the perceptual defense
model has been that the phenomena in question could be
described in a more parsimonious fashion; that "Response
Biasing"--the frequency of certain words or verbal associ
ations in the subjectif social environment— dictates how
he will respond to these verbal stimuli (Minard, 1965).
For example, mediated generalization theory suggests that
when presented with the word pair "BUTTON-BREAD" the impli
cit response "white" to the explicit word "BREAD" causes
the word "white" to present itself in contiguity with the
word "BUTTON" in the mind of the subject. Thus, strength
ening the "BREAD-white" association and increasing the
likelihood that the subject would readily attain the con
cept "white" in the subsequent list. According to Minard
(1965) this assumes that the frequency of the implicit
response "white" to the explicit word "BREAD" is the same
for all groups of people. Applying this argument to the
present study, the equality of the strength of certain
word associations for both black and white subjects could
be questioned. Are the response hierarchies for a given
verbal stimulus similar for the two groups? The hypothe
sis set forth in this paper predicted that if one equated
vocabulary scores of blacks and whites one could conse
quently assume similar verbal associations, since common
and equivalent definitions found in the dictionary would
have to be given for maximum scoring. Too, when "soft"
79
was the mediated concept, one could predict performance
under the mediated condition. These findings of the
present study, therefore, are inconclusive in answering
the question of a difference in response hierarchies.
Limitations and Implications
for Future Research
The present study was founded upon two basic as
sumptions which results suggest may need additional inves
tigation.
1) Response hierarchies for a given verbal stimulus are
the same for black and white students if one equates IQ
scores of the two groups. In an attempt to replicate prior
investigations as precisely as possible (Jacobson, Millham
and Berger, 1969; Millham, Jacobson and Berger, 1971), the
present study did not research the response hierarchies
of the words employed for black college students. It
would be worthwhile to calculate the response dominance of
particular words for both black and white subjects and
proceed to analyze the similarities and dissimalarities in
verbal conotations.
2) Vocabulary IQ scores are excellent predictors of Full
Scale IQ scores for black students as they are for white
students. Taking into consideration the issue of "cultural"
80
IQ test biasing, especially on verbally laden items, it
is conceivable to this researcher that a low vocabulary
score for a black student does not suggest a low Full
Scale IQ score or at least, not low level intellectual
functioning.
Another limitation of the present investigation is
its inability to access the effects the experimenters sex
had on these students. While previous investigations em
ployed male examiners, the present study utilized female
examiners. Do students perform better when they are tested
by a same sex examiner, by an examiner of the opposite sex,
or does sex of the examiner have no significant effect on
the performance of subjects under the conditions of studies
similar to the present one?
Inspite of the limitations put forth, the present
research did indicate that the effects of intellectual
functioning and mediation on conceptual learning are the
same for black college students as they are for white
college students when appropriate stimuli are employed.
If the results of this study are generalizable, the utili
zation of certain concepts (e.g., "white" for black college
students) may be inappropriate for a given population due
to the stimuli's anxiety arousing effects. The implica
tions of this finding for future studies of this type is
far-reaching. It would be paramount that researcher
evaluate the social values of intended stimuli prior to
81
the collection of data and the evaluation of the results.
It is possible that investigators who have not controlled
for this phenomenon in conceptual learning studies have
not measured a subject’s ability to utilize mediators or
to conceptualize, but in actuality measured the level of a
subject's defensiveness or defensive reactions to social
(verbal) stimuli; and just as misinterpretations have
been made in the past, they will continue to be made in the
future unless researchers are sensitive to the issues in
volved and consider social values in verbal learning.
In addition, it was demonstrated that lower IQ
scoring black students' academic performance was at the
same level as the higher scoring groups'. In the wake of
the present legal and political unrest surrounding the
belief that receiving a certain score (high score) on a
standardized test indicates that you are "more qualified"
for admissions to certain programs (e.g., college, medical,
law, or graduate schools), than someone who scored lower,
the results of this study call into question the generali-
zability of that belief to black students which could have
a far-reaching implication if replicated.
82
REFERENCES
83
References
Bourne, L. E., Jr. Human Conceptual Behavior. Boston:
Allyn and Bacon, 19 66.
Brown, W. P. Conceptions of perceptual defense. British
Journal of Psychology, Monograph Supplement, 19 61,
No. 35.
Coleman, E. G. Verbal concept learning as a function of .
instructions and dominance level. Journal of
Experimental Psychology, 1964, 213-214.
Clyde, D. J., Cramer, E. M, and Sherin, R. J. Multivar
iate statistical program. Coral Gables, Florida:
Biometric Laboratory, University of Miami, 196 6.
Friendly. F. and Bercrer. S. Intellectual functionincr and
mediation in black and white first grade children.
Presented at the California State Psychological
Association Convention, March, 1976.
Jacobson, L. I., Dickinson, T. C., Fleishman, J. M., and
Haraguchi, R. S. The relationship of intelligence
and mediating processes to concept learning.
Journal of Educational Psychology, 1969, 60, 109-
112.
Jacobson, L. I., Elenewski, J. J., and Lordahl, D. S. The
relationship between rote learning and mediated and
nonmediated concept learning. Psychonomic
Science, 1968, 347-348.
Jacobson, L. I., Elenewski, J. J. Lordahl, D. S., and
Liroff, J. H. Role of creativity and intelligence
in conceptualization. Journal of Personality and
and Social Psychology, 1968, ]^, 431-436.
Jacobson, L. I., Millham, J., and Berger, S. E. Individual
difference in information processing during con
cept learning. Psychonomic Science, 1969, 14,
287-289.
Kendler, T. S., and Kendler, H. H. Reversal and non-re
versal shifts in kindergarten children. Journal
of Experimental Psychology, 1959, 58 56-60.
84
Jensen, A. R. How much can we boost IQ? Harvard Educa
tional Review, 1969, 39(1), 1-123.
Jensen, A. R. Genetics and Education. San Francisco;
Harper and Row Publishers. 19 72.
Jensen, A. R., and Rohwer, W. D., Jr. Experimental analy
sis of learning abilities in culturally disad
vantaged children. Final Report on OEO Project
No. 2404, U. W. Office of Economic Opportunity,
1969.
Johnson, R. C., Thompson, C. and Frincke, G. Word values,
word frequency and visual deviation thresholds.
Psychological Review, 1960, ^1_ 332-342.
McGinnis, E. Emotionality and Perception Defense. Psy
chological Review, 1949 , 244-251.
Mednick, S. A., and Freedman, J. L. Facilitation of con
cept formation through mediated generalization.
Journal of Experimental Psychology, 196 0, 60,
278-283.
Minard, J. G. Response-bias interpretation of "Perceptual
Defense." Psychological Review, 1965, 7_2, 74-88.
Millham, J., Jacobson, L. I., and Berger, S. Effects of
intelligence, information processing, and mediation
conditions on conceptual learning. Journal of
Educational Psychology, 1971, 62 (4) , 293-299.
Osier, S. F., and Fivel, M. W. Concept attainment: I.
The role of age and intelligence in concept at
tainment by induction. Journal of Experimental
Psychology, 19 61, 1”^*
Osier, S. F., and Trautman, G. E. Concept attainment : II.
Effects of stimulus complexity upon concept at
tainment at two levels of intelligence.
Journal of Experimental Psychology, 19 61, 6^, 9-13.
Osier, S. F. , and Weiss, S. R. Studies in concept at
tainment : III. Effects of instructions at two
levels of intelligence. Journal of Experimental
Psychology, 1962, 62 y 528-533.
Russell, W. A,, and Jenkins, J. J. The complete Minnesota
norms for responses to 100 words from the Kent-
Rosanoff Word Association Test. ONR tech Rep.,
1954, No. 11.
85
Shockley, W. Negro IQ deficit: Failure of a "malicious
consequence" model warrants new research proposal.
Review of Educational Research, 19 71, r 22 7-248.
Swinger, H., Berger, S., and Friendly, F. Race and Motive
to avoid success. Presented at the American Psy
chological Convention, September, 19 76.
Underwood, B. J., and Richardson, J. Some verbal materials
for the study of concept formation. Psychological
Bulletin, 1955, 53, 84-95.
Vinacke, W. E. The Psychology of Thinking. New York:
McGraw-Hill, 1952.
86
FAMILY BACKGROUND QUESTIONNAIRE
FAMILY BACKGROUND QUESTIONNAIRE
NAME
GRADE IN SCHOOL:
MAJOR:
GRADE-POINT-AVERAGE :
NUMBER OF CHILDREN IN FAMILY
ELEMENTARY SCHOOL Jr.High
MOTHER'S LAST GRADE COMPLETED
(Check One)
(1) (2) (3) (4) (5) (6) ■ (7) (8) (9)
ADVANCED COLLEGE
HIGH SCHOOL (Graduate School)
(10) (11) (12)
COLLEGE
(13)(14)(15)(16)
(17+)
MOTHER'S OCCUPATICN:
ELEMENTARY SCHOOL Jr.High
FATHER' S LAST GRADE COMPLETED: (1) (2) (3) (4) (5) (6) (7) (8) (9)
(Check One) ADVANCED COLLEGE
HIGH SCHOOL (Graduate School)
(10) (11) (12) (17+)
COLLEGE
(13)(14)(15)(16)
FATHER'S OCCUPATION:
FAMILY INCOME (strictly confidential)
88

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Negro participation in civil government with emphasis on public education in Texas

PDF

Dual devotions: African American clergywomen and work-family dilemmas

PDF

Intellectual functioning and mediation in white and black children

PDF

The 'Widmar' (1981) decision and the status of religious liberty on the public college and university campus: Antecedents, effects, and prospects

PDF

The efficacy of hypnotic analgesia for dental pain in individuals of both high and low hypnotic susceptibility

PDF

An experimental study of the effect of aikido training on the self-concept of adolescents with behavioral problems

PDF

The relationship between dual-earner couples' daily work mood and their daily marital satisfaction and home affect

PDF

A national study of Title IX and its impact upon public education in the United States from 1972 to 1980

PDF

Contracting with private agencies for the delivery of public services: Policy implications on an emerging trend

PDF

Dental health services and the aging: a gerontological perspective for social policy

PDF

Site selection and laser properties of neodymium and europium ions in fluorapatite laser crystals

PDF

The effects of violating selected item-writing principles on the difficulty and reliability of multiple-choice tests for health profession students

PDF

The cross-border metropolis hypothesis: Intra-urban structure of neighboring cities, Tijuana (Mexico) and San Diego (USA)

PDF

Dental care of the aged patient: Implications of attitudes toward aging among dental students and dental hygiene students

PDF

Women in the skilled crafts: predictors of nontraditional job choice

PDF

The use of cubic splines for estimating the prognostic effect of age at diagnosis in childhood acute lymphoblastic leukemia

PDF

An evaluation of an innovative master's degree program for administrators of ambulatory health care centers

PDF

Role accommodation: A study of optometrists in a prepaid medical group as perspective for viewing relations between two professiosl occupations

PDF

The role of the sensorimotor cortical system in skill acquisition and motor learning: a behavioral study

PDF

The economics of tenure: understanding the effects of ethnicity, status and discipline on faculty attitude, workload and productivity

Asset Metadata

Creator Friendly, Felicia (author)

Core Title The effects of intellectual functioning and mediation on conceptual learning for Black college students

School Graduate School

Degree Doctor of Philosophy

Degree Program Psychology

Degree Conferral Date 1978-05

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

Tag education,OAI-PMH Harvest,social sciences

Format application/pdf (imt)

Language English

Contributor Digitized by ProQuest (provenance)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c37-69734

Unique identifier UC11632812

Identifier DP71535.pdf (filename),usctheses-c37-69734 (legacy record id)

Legacy Identifier DP71535.pdf

Dmrecord 69734

Document Type Dissertation

Format application/pdf (imt)

Rights Friendly, Felicia

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