Canaries in the coal mine: Perceptions of threat and media system dependency relations.

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content CANARIES IN THE COAL MINE:
PERCEPTIONS OF THREAT AND
MEDIA SYSTEM DEPENDENCY RELATIONS
by
William Earl Loges
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Communication Theory and Research)
May, 1992
Copyright 1992 William E. Loges
UMI Number: DP22477
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.
Dissertation Publishing
UMI DP22477
Published by ProQuest LLC (2014). 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
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, Ml 48106 -1346
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANOBLBS, CALIFORNIA 90007
This dissertation, written by
under the direction of f t . . i s Dissertation
Committee, and approved by all its members,
has been presented to and accepted by The
Graduate School, in partial fulfillm ent of re
quirements fo r the degree of
CH
'°}2
L. 8 3 2 .
DOCTOR OF PHILOSOPHY
Dean of Graduate Studies
DISSERT
Chairperson
Dedicated to
Beverly J. Loges,
without whom it wouldn't have been possible,
and to
Jessica R. Armstead-Loges,
without whom it wouldn't have been worthwhile.
• t I
111
ACKNOWLEDGEMENTS
In the last five years, I've been forced to
^operationally define the concept "support" in a multitude
!of ways. No matter which way I've had to seek support,
Sandra Ball-Rokeach has provided it. Nevertheless, as far
ks I can discern, she has spared me no criticism, and this
(document has benefitted at every stage from her generosity.
Other faculty members at the Annenberg School for
jCommunication have been instrumental in the preparation of
this dissertation. I gratefully recognize the efforts of
Jim Beniger, particularly for his advice regarding
'statistical analysis. Peter Clarke and Susan Evans
provided important motivation as I began drafting the first
!
^chapters. As Dean of the School, Peter Clarke also
provided administrative support for the research project
which yielded the data analyzed in the pages to follow. In
i
I '
the department of Communication Arts and Sciences at the
[University of Southern California, Karen Altman patiently
I
^tolerated my sudden appearances with drafts of lengthy
(
[documents at various stages of my graduate education, and
never failed to provide encouragement and excellent advice.
! Several staff members at the Annenberg School have
earned distinguished mention in more than one previous
dissertation.^For their ceaseless material and spiritual
support Agnes Uy, Pat Boyd, Cecilia Pastrana, Valerie Hunt,
‘ Anne-Marie Campian, and Linda Caban all have my gratitude
and enduring respect. The corps of work study students who
Jcopied thousands of pages, fetched dozens of books and
i
[journals, and put up with any amount of abuse as I
f
nervously raced around the CommCenter (the basement library
jof the Annenberg School) are an amazing bunch. They have
jcome and gone in the five years that I stayed, and they
»
t
jwere always terrific.
I
I
j The Media System Dependency Research Group, whose
i
collective efforts led to the development of the survey
instrument and data collection described in Chapter Four,
^were tireless and dedicated. Besides the group's anchor,
I
Sandra Ball-Rokeach, they include Dr. Augie Grant, Dr.
Carolyn Aydin, Dr. Gerry Power, Dr. Kendall Guthrie, Ross
I
Waring, Lynn Dickinson, August Horvath, and Kelly Madison.
Thanks too to Dr. Wendy Colman.
The community of students at the Annenberg School with
jWhom I have worked, drank, and played (not always in that
(
,order) has been wonderful. Those who entered the program
jwith me, and who have made this journey bearable in the
I
strange years that followed our orientation meeting in
I
August of 1986, are cherished friends. Tom Valente, who
arrived the next year, and Rebecca Davis, who never arrived
officially but is an honorary Annenberger, have been
incomparable friends. I highly recommend them to anyone in
graduate school. This dissertation gained momentum after aj
crucial conversation with Brian Boyd in the spring of 1990,
I
land his contribution to clarifying my thoughts on these '
subjects is much appreciated. Two students who tolerated !
I
my agonies in the Stat Lab with equanimity and good humor !
I
jare Lori Collins-Jarvis and Jennifer Monahan.
All of my acquaintances knew I was writing a
i
dissertation (or, at times, knew that I was not writing a
;
jdissertation). Those friends who constantly provided
encouragement, comfort, affection, and free meals include
Willie Ambrico (a room mate without equal), Judy Morgan
(who had a zero-tolerance policy for my self doubts), Diane!
O'Dea (who provided inspiration and energy at every crucial
I 1
j
point), Paul Krask, Pat Oliver, Brian and Maria Allen, Lena
I
jChao, Doug Shook, Rhonda Shook, Frank Webster and Liz
I
jChapman. Perhaps without knowing it, each of these people
■kept me at this long after I thought it was worth while,
i My family has played an enormous role in my graduate
education. My youngest brother, Keith, and my oldest
i
jsister, Dawn, each deserve special thanks for their
?unfailing support.
i
; Special recognition is due as well to my employers at
I
Offisys in Burbank and Woodland Hills for their patience
with my awkward schedule and priorities in the last half of
,1991.
In the peculiar economy of ideas and friendship, the i
i
I
debts that I owe to the people listed above will largely be
paid through the efforts I make on behalf of students and
colleagues in the years to come. I cheerfully accept the
> i
1 . i
burden of such an obligation if I may look forward to the !
I i
continued company of such wonderful people as these.
With the benefit of such a special group to advise and
'inspire me, it may be surprising that this document is
•flawed at all. It will suffice to say that I have managed
I
to assert my shortcomings even in the face of the 1
collective efforts of these precious people. In the
1 ' !
laundry room of the apartment complex in which I live there!
is a sign which reads "User must assume responsibility for
results!" While my landlord may intend for the sign to
refer only to laundry, the sentiment is worth keeping in
i
mind where one's education is concerned. At any rate, I do
i
not seek by naming names to avoid responsibility for the
i
weaknesses in the pages to follow. If you, reader, are
!
baffled by the lapses to come, you should have seen the
drafts that Profs. Ball-Rokeach, Beniger, and Altman
l
endured and improved . . .
I
William E. Loges
North Hollywood, CA
September^, 1991
i
i
vii
TABLE OF CONTENTS
Page
Acknowledgements i i i
List of Tables ix
Abstract xiii
Chapter One: Media, Information, and Threat 1
The Social Role of Information Systems 3
Media System Dependency Theory 8
Information Seeking and Threat 25
Chapter Two: The Concept of Threat 34
Threat and Stress 39
Threat and Risk 47
Threat and Ambiguity 54
Summary 64
Chapter Three: Research Questions and Hypotheses 65
Chapter Four: Methods 71
The Development of a Media System Dependency
Scale 71
A Scale to Measure Perceived Threat 76
Measures of Media Use 77
Other Items on the Survey Instrument 78
The Instrument in Two Communities 80
The Sample and Design 83
Analyzing the Data * 96
Chapter Five: Results 102
Threat Perceptions 102
Media System Dependency Relations 114
Threat Perceptions and the Intensity and
Substantive Scope of Dependency Relations 134
Threat Perceptions and the Media Scope of
Dependency Relations 148
Threat Perceptions, Education, Income, and Age 157
Media Use, Threat, and Dependency Relations 162
Summary 174
Chapter Six: Discussion 178
Threat Perceptions and Dependency Relations 178
A Definition of Threat 185
Threat Perceptions and the Scope of Dependency
Relations 186
viii
Threat Perceptions and Media Use 190
Dependency Relations With Newspapers 193
Dependency Relations With Television 195
Some Anomalous Findings for Radio and Magazines 197
Weaknesses in the Present Study 206
Suggestions for Further Research 209
Conclusion 215
References 219
Appendix A: The San Bernardino Community Survey 226
Appendix B: The Austin Community Survey 237
LIST OF TABLES
Table 4.1
Table 4.2
Table 4.3
Table 4.4
Table 4.5
Table 4.6
Table 5.1
Table 5.2
Table 5.3
Table 5.4
Table 5.5
Table 5.6
Response rates in Austin and San
Bernardino
Summary statistics for age, sex, and
education, for respondents from Austin
and San Bernardino
Summary statistics for years of residence
in the community, ethnicity, and
employment status, for respondents
from Austin and San Bernardino
Summary statistics for income for
respondents from Austin and San
Bernardino
Summary statistics for age, sex,
education, and years of residence in
the community for the combined data
set
Summary statistics for ethnicity,
employment status, and income for the
combined data set
Reliability analysis of Threat Perception
Scale
Bivariate correlations between threat
perceptions and age, education, and
years of residence in the community
Bivariate analysis of threat perceptions
and sex. Analysis of variance.
Bivariate analysis of threat perceptions
by income. Scheffe test.
Bivariate analysis of threat perceptions
by employment status. Scheffe test.
Bivariate analysis of threat perceptions
by ethnicity. Scheffe test.
Page
87
91
92
93
94
95
103
106
108
109
111
112
Table 5.7 Reliability analysis of the individual
level Media System Dependency Scale.
Cronbach's alpha.
X
118
Table 5.8 Factor analysis of the individual level
Media System Dependency Scale for
newspapers. Oblique rotation. 122
Table 5.9 Factor analysis of the individual level
Media System Dependency Scale for
radio. Oblique rotation. 123
Table 5.10 Factor analysis of the individual level
Media System Dependency Scale for
magazines. Oblique rotation. 125
Table 5.11 Factor analysis of the individual level
Media System Dependency Scale for
television. Oblique rotation. 127
Table 5.12 Regression Analysis of threat perceptions
and media system dependency relations
for newspapers 135
Table 5.13 Analysis of variance between threat
perceptions and the substantive scope
of media system dependency relations
for newspapers. Scheffe test. 137
Table 5.14 Regression analysis of threat perceptions
and media system dependency relations
for radio 139
Table 5.15 Analysis of variance between threat
perceptions and the substantive scope
of media system dependency relations
for radio. Scheffe test. 141
Table 5.16 Regression analysis of threat perceptions
and media system dependency relations
for magazines 142
Table 5.17 Analysis of variance between threat
perceptions and the substantive scope
of media system dependency relations
for magazines. Scheffe test. 144
Table 5.18 Regression analysis of threat perceptions
and media system dependency relations
for television 145
xi
Table 5.19 Analysis of variance between threat
perceptions and the substantive scope
of media system dependency relations
for television. Scheffe test. 146
Table 5.20 Analysis of variance between threat
perceptions and the media scope of
media system dependency relations for
social understanding. Scheffe test. 149
Table 5.21 Analysis of variance between threat
perceptions and the media scope of
media system dependency relations for
self understanding. Scheffe test. 150
Table 5.22 Analysis of variance between threat
perceptions and the media scope of
media system dependency relations for
interaction orientation. Scheffe
test. 152
Table 5.23 Analysis of variance between threat
perceptions and the media scope of
media system dependency relations for
action orientation. Scheffe test. 153
Table 5.24 Analysis of variance between threat
perceptions and the media scope of
media system dependency relations for
social play. Scheffe test. 154
Table 5.25 Analysis of variance between threat
perceptions and the media scope of
media system dependency relations for
solitary play. Scheffe test. 156
Table 5.26 Regression analysis of newspaper use and
threat perceptions, media system
dependency relations, demographic
variables, and use of other media. 164
Table 5.27 Regression analysis of radio use and
threat perceptions, media system
dependency relations, demographic
variables, and use of other media. 167
Table 5.28 Regression analysis of magazine use and
threat perceptions, media system
dependency relations, demographic
variables, and use of other media. 169
Table 5.29 Regression analysis of television use and
threat perceptions, media system
dependency relations, demographic
variables, and use of other media.
xii
171
xiii
ABSTRACT
In her statement of the propositions of Media System
Dependency Theory, Ball-Rokeach proposed that one's
perceptions of increasing levels of threat in his or her
personal and social environment would lead to the
intensification of media system dependency relations.
This proposition was not tested in prior research. While
the proposition is consistent with other mass media
theories, it is not entirely consistent with the results
of research into the interpersonal communication behavior
of individuals under stress and of corporate executives in
uncertain business environments. This dissertation
proposes a comprehensive definition of "threat" and
proceeds to hypotheses testing Ball-Rokeach's proposition
regarding the correlation between threat perceptions and
media system dependency relations. Survey data from two
cities are used, and the results of a multiple regression
analysis support Ball-Rokeach's claim that those who
perceive greater levels of threat from crime, health
concerns, economic recession, and ecological or natural
hazards have more intense media system dependency
relations than those who perceive less threat from their
environment.
i:
Chapter One: Media, Information, and Threat |
!
It appears that the relationship between a person's '
l
perception of the environment as threatening and the \
| t
association of that perception with patterns of media use |
I
jis in need of more precise articulation and study. In the j
! . !
United States, crises are expected to produce massive
i
audiences for the mass media. The attention Americans paid)
to television in the days following the assassination of
i
President Kennedy are often used as an example of this
i
phenomenon, and the Gulf War of 1991 produced a similar
I
reaction. The desire to acquire information in the face of
jan acute threat and the belief that the mass media are
helpful sources of pertinent information, seem so self-
i
[explanatory that they need no further study.
I
i
I Is it safe to presume that the desire for more
i
linformation extends to chronic threats? Is it always
considered better to have more information, even when a
threat is perceived to be part of daily life? Studies of
i
patients awaiting surgery (Janis, 1958), of people living
'in environmentally hazardous areas (Sandman, 1988), and of
business executives facing uncertain economic conditions
I
[(Boyd, 1989) have suggested that, at least for some people,
it is considered best not to seek information, or at least
not to expand one's search for information, in reaction to
phese dangers. Is this a reasonable way to behave in
response to a threat?
| Two distinct models of information-seeking in response
,to threats are suggested. Ball-Rokeach (Ball-Rokeach,
Rokeach, and Grube, 1984; Ball-Rokeach, 1985) has proposed
jthat the intensity of media system dependency relations
will increase as the amount of threat one perceives in the
|
environment increases. This observation is based on the
I
proposition that among the functions of the mass media
jsystem is to provide informational resources which assist
people in reaching goals of understanding, orientation, and
I
play. Perceived threats make these goals harder to attain,
i
and the resources of the media then become more prized.
®all-Rokeach offered no theoretical limit for the increase
I
(in intensity, and thus described a linear model, with
i
perceptions of threat on the horizontal axis and the
l
I
intensity of dependency relations on the vertical axis.
Ball-Rokeach describes the relationship between mass media
and people who perceive high levels of threat like the
i
relationship between a canary and a coal miner. The more
i
poisonous the miner believes the environment to be, the
more helpful the canary is perceived to be.
I Conversely, Simon's (1955) concept of "satisficing"—
the cessation of information-seeking when its costs exceed
|
its benefits—suggests that people may stop their search for
jinformation when the psychological toll of the information
they get starts to exceed their tolerance. People have
boundaries placed on the rationality they can afford to
■apply to their decision-making, and some of those
boundaries are psychological. Simon's model is may be
graphed as a convex parabola, where information-seeking is
on the horizontal axis and the cost of information is on
the vertical axis.
This chapter proceeds to a discussion of mass media
I
theory and research as it has pertained to threat—
l
particularly media system dependency theory. I then
Jdiscuss Simon's ideas in more detail, and the work of
Jcertain organizational scholars concerned with issues
related to threat perceptions and information-seeking. In
subsequent chapters I discuss the concept of threat in
finer detail, present hypotheses, and describe the research
method used to test them. I conclude with the results of
I
jthese tests and a discussion of the results.
! The Social Role of Information Systems
In a simple society, information about one's social
environment—including one's economic security, personal
health, and safety from crime and natural disasters—can be
!
i
,obtained in a relatively direct fashion from priests,
jchief tains, doctors, and elders without professional media
I
spokespeople. Complex economic and political systems tend
to develop centralized and distinct media systems that are
'responsible for creating, gathering, processing, and
i
Jdisseminating information (Ball-Rokeach, 1985).
Technologies that allow the processing of information
jalone, such as the telegraph and radio, allow the media
■system to distinguish itself by a mode of production that
does not rely upon the society's general transportation
system. It is control of these technologies and other
i
resources (resources as diverse as forests, satellites, and
the labor of writers) that allows a media system to
accomplish the functions of creating, gathering,
processing, and disseminating information. The variety of
means of control includes outright ownership, government
jlicense, and contractual agreement.
j Some societies seek to minimize the diffentiation
!
between the media system and the government, such as the
i
U.S.S.R. before perestroika. while some seek to maximize
that separation, such as the jealously guarded press
i
freedom in the United States. The media system is defined
by its functions and by its specific forms of control over
i
|the resources needed to perform those functions (Ball-
i
iRokeach, 1985) .
I
I
j The informational goals of most people in the United
States can be served by the major mass media more
thoroughly and more conveniently than by any religious,
I political, or social group. News, weather, sports, j
I '
entertainment, health, fashion, food, and commercial
f
announcements are routinely the subjects of television
i
programming each day. They make up the contents of most j
major newspapers. It is not surprising that other social ’
j
{systems, such as the government and industry, have come to ‘
frely upon the informational resources of the mass media
[system for creating, gathering, processing, or |
I >
disseminating political and commercial information. These ]
other social forces recognize that people not only go where
information is, but that people go where.they have come to !
expect information to be.
The mass media become a link from the person to the
(larger social structures from which the person requires
information. How secure is one's job? Major employers may
lay off thousands of employees with little or no warning.
The prognostications of economic forecasters on the evening
|news, or of local reporters in search of a big story, may
[prove to be better signals of job security than the
^assurances of one's boss. How sound is one's health?
jPeople complain that their doctors speak to them in jargon,
i
jmystifying their complaints and cloaking the cure in
jsecrecy (Evans and Clarke, 1983). Popular medical advice
i
(programs and personalities encourage people to be aware of
~ 6 [
their own health habits and of the precautions against
illness they can take.
i In a social environment made up of moderately
|
juncertain conditions, these informational resources can
l
ja source of reassurance and stability for most people.
|But when conditions in one's social life are perceived to j
|be threatening, and one expects or fears economic hardship,1
-illness, crime, or natural disaster, the mass media may be J
of less comfort. That is not to say that one needs less
!
1 • . . 1
; information of the sort offered by the mass media under
i
I ;
conditions of increased threat, but to say that the mass
I !
media may not be perceived to be the most helpful sources j
of such information. Also, if the perceived threat is
i
grave enough, one may not consider the acquisition of more
i !
information to be useful at all. One may decide that
I •
I
action is needed which need not or can not involve
i
monitoring the wider environment through media use.
Coherent theories of how mass media are used and with
l
Jwhat effects have generally included linear relationships
between exposure, certain specified social or psychological
attributes, and certain effects. While many of the most
widely researched perspectives on mass media theory, such
as uses and gratifications, agenda setting, and cultivation
analysis, offer complex conceptual relationships, reviews
I
of their research literature tend to reveal a relatively
narrow range of linear relationships under study
(Palmgreen, 1984; Rogers and Dearing, 1988). Increases in
juse are hypothesized to lead to enculteration, or adoption
jof the media agenda, or gratification of one or more
\psychological or practical needs.
i
Hypotheses regarding the relationship between media
use and the desire to reduce or at least to be fully aware
of the level of threat in one's environment have generally
predicted a linear relationship. Increased media use is
jhypothesized to be correlated with an increase or a
decrease in perceived threat. Gerbner et al. demonstrate
i
the former through cultivation analyses (Gerbner and Gross,
1976; Gerbner, Morgan, and Signorielli, 1980); certain uses
l
and gratifications research demonstrates the latter (Katz,
|Blumler, and Gurevitch, 1974; Lain, 19861) . Conversely,
increases in perceived threat are hypothesized to be
jcorrelated with increased media use (Katz, Gurevitch, and
iHaas, 1973).
■ One way to address the question of the impact of media
!
use is to examine the concept of "use" in a more complex
manner. While it may be banal to assert that not all use
j 1Lain's study of newspaper reading attempts a more
complex theoretical approach than much of the previous work
in newspaper use (c.f. Bogart, 1981 for a review of
newspaper studies). Lain concludes that "[njewspapers
ought to promote themselves as the repositories of the
information that will relieve anxieties and hasten
integration into the community" (Lain, 1986, p. 74).
of the media is the same, the challenge is to account for j
i
the difference in a reliable and theoretically parsimoniousl
i
fashion. One dimension of use is the expectation of a !
"user" that the media will provide information that can be j
I helpful in the pursuit of one or more central personal j
| ;
goals. The impact of these expectations on habits and
{beliefs has been examined using media system dependency
theory.
Media System Dependency Theory
Media system dependency theory (MSD) offers a complex i
i
conceptualization of the relations in which individuals, 1
groups, organizations, and systems are engaged with the
i ■
media system. A central proposition in the definition of !
the relations between the media system and individuals is
that the relations are asymmetrical (Ball-Rokeach et al..
1984; Ball-Rokeach, 1985; Emerson, 1962). This means that
jthe informational resources controlled by the media system,
i
and desired by individuals for pursuit of their goals, are
'more important than the resources controlled by any one
individual, and desired by the media system for pursuit of
its goals (Ball-Rokeach, Power, Guthrie, and Waring, 1990).
At higher levels of social structure, such as groups,
jorganizations, and social systems, the relations of
dependency are not as presumptively one-sided in favor of
the media system. The need of the media system for
I information, for instance, may provide information-rich
I
organizations (such as the Pentagon during wartime) with
I
the balance of power in the relation (Ball-Rokeach, 1985;
Ball-Rokeach et al.. 1990). 1
I
j While the structure of media system dependencies at !
| j
jthe individual level is regarded as asymmetrical, the ■
* I
I . I
intensity and scope of MSD relations at the individual !
i ;
;level are variable. The intensity of a media system j
f I
(dependency relation is the measure of the asymmetry—the
iextent to which the media system's informational resources
I
jare perceived by an individual to be particularly helpful
j
I in the pursuit of his or her goals. Intensity will vary as
f
'goals vary, as personal and social environments vary, as
i
perceptions of the media system vary, as access to the
:media system varies, and as the exclusivity of the media
system's control over resources varies (Ball-Rokeach,
! 1985) .
i
j The scope of media system dependency relations is
■conceptualized as the range of resources, media, and
i
substantive motivations implicated in an individual's media
system dependency relations. Of the types of informational
.resources controlled by the mass media (resources for
■creating, gathering, processing, and disseminating
information), most people most times require the
f
dissemination resources for their individual purposes. It
is the rare case, in other words, when the goals of an
individual specifically implicate the information
^processing resources of a mass medium (for instance, one
may never find it necessary or helpful to personally use a
videotape editor). Further, it is rare that individuals
have access to the creating, gathering, and processing
resources of the media, even if they wish to use these
!resources. It is far easier to obtain a newspaper than a
[printing press. For all these reasons, the resource scope
of individual level dependency relations is usually
narrowed to the dissemination resources.
Media scope can vary quite dramatically, however,
because it refers to the range of media forms a person may
;find helpful in the pursuit of goals. Media forms can be
1
conceived at broad levels—broadcast, electronic, or print
forms. Media forms can also be conceived at more specific
levels—e.g. genre or specific technological forms (such as
, television). Media scope is not measured by counting the
j
; hours devoted to using each media form but by measuring the
i
i
jintensity of relations across a variety of forms. Media
!
jscope may vary as a result of the perceptions of assorted
|forms individuals develop after exposure to the forms, or
jas a result of access to a limited range of forms.
! Substantive scope is the range of types of goals a
person pursues in his or her relations with the resources
ini
I
of the media system. Goals of understanding, orientation, \
and play are conceived to be the basic types of individual
substantive motivations for media relations. The types of
goals are further conceptualized at the individual and
!social levels. Understanding is specified as self !
understanding and social understanding. Orientation is i
specified as action orientation and interaction orientation,
| ;
j(i.e. orientation to social interaction). Play is
I
i
ispecified as solitary play and social play. These types
;are conceived to be exhaustive of the universe of media- :
j
!related goals at the individual level, but not mutually
|exclusive. That is, media resources may be desired in one ;
{instance to pursue two different goals. People's
I
expectations regarding the goals that a television program
or newspaper article will address (e.g. goals of solitary
play or social understanding) vary, and these expectations
|
help determine the level of exposure a person will decide
i ,
jto engage in, and the level of effect the media may have in
iany given instance (Ball-Rokeach et al.. 1984; Colman,
! 1990).
Self understanding is the development of a self
.
tconcept that includes self-esteem, beliefs, values, and
I
iattitudes. While intensely personal, self understanding is
i
ja goal that may certainly be pursued through the
i
! acquisition of information the mass media offer. For
instance, radio talk shows feature the advice of
psychologists on a variety of subjects that may help a
person grapple with an array of personal problems. While
the context of the show is that of a broadcast aimed at
ibroad masses, the informational content is deliberately
iintimate, and can be accepted as pertinent to oneself '
regardless of the size of the audience. !
Social understanding is the knowledge one has of how
society and its institutions function, and the perception
i
jone has of his or her role in that society and in those
institutions. While at any single moment a person may feel
[comfortable with his or her understanding of these matters,
human growth and the changes in society—the changing roles
jwe play and the changes in the expectations in those roles—,
j
jmake social understanding an ongoing project for most
jpeople. Mass media offer an array of resources for
! pursuing social understanding. For instance, news
magazines provide current information about the activities
of government institutions, the economic system, the arts
and entertainment, and other information that is considered
j !
by the editors to be pertinent at a level of general social
[acquire, over time, a generalized concept of the
relationship between social systems (e.g. between the
|interest. Readers gain not only the Mfacts," but a sense
i
iof what society considers important. They may also
13
government and industry, or between the media system and
the rest of the economy).
Action orientation is the goal of behaving in a way
t
ithat is consistent with the expectations and norms of a
society in a given situation. People seek to orient their
actions such that it is possible to perform the action
effectively and still not give offense to prevailing social
|norms. It cannot be assumed that people even know
[intuitively the range of actions available to them.
t
Gaining information that provides action orientation is as
much a matter of learning what may be done as how it should
be done. Information that helps one orient his or her
action is available in the media. An example is local
television news. The diversity of stories on a local
newscast will generally include topics such as crime
(acting legally), consumer advice (acting responsibly), and
human interest stories (acting nobly or humanely). More
[specifically, one may learn what commodities are available
i
|for purchase, and where, and one may learn details about
upcoming elections. In short, not only may one's search
for a general orientation to action be facilitated through
[the resources of the mass media, but one's orientation to
;specific actions, such as shopping and voting, may be
!
facilitated through the same resources.
3. 4;
I
I
Interaction orientation is the acquisition and
development of suitable and effective conversational,
{social, and other interpersonal manners. Knowing how to >
|
address one's elders, one's employers, and one's ministers I
1 t
{demonstrates an ability to orient one's interaction in a
jmanner appropriate to one's social role, its attendant
I I
(expectations, and one's personal goals. Also, knowing how i
I ;
I to respond to the addresses of salespeople, beggars, and .
I . :
{children demonstrates awareness of one's status in society.'
i
Not all of these interaction orientations are learned j
through experience. Some, such as addressing royalty with j
i
I
a bow or curtsey, must be acquired second-hand, and some of!
jthat learning will be done through mass media. Newspaper I
i
.advice columns, for instance, often tell young people the
proper wedding etiquette. They also inform people about
race relations, family relations, and many other matters
involving human social interaction that their
correspondents, for whatever reasons, have found
troublesome or ambiguous.
Solitary play is the entertainment or diversion of
one's self from the more demanding and stressful parts of
i
i
life. This is not to minimize the importance of play. The
I
lability of any mechanism to provide solitary play gains for
t
|the device an immediate interest, from home computers to
rag dolls. The mass media, particularly radio and
I
television, are clearly important for their entertainment 1
I
resources. It is to these resources that most of j
broadcasting's investment is directed; the purchase of j
programs and the scheduling of entertainment programming in;
the most heavily populated viewing or listening hours j
(prime time for television, drive time for radio). The
activities that provide solitary play are not the same for ■
i
ieveryone, though, and that allows people with various j
{interests to define reading the newspaper or technical
journals as a sort of playful diversion while others may
I consider reading such information to be stressful or
1
idifficult. Playing is not defined as "watching sitcoms,"
I
.but as relaxing, unwinding, or diverting oneself from cares
|regardless of the means through which this is accomplished.
I
Social play is entertainment or escape in and through
the company of others. A dinner party, a bowling league,
or a high school dance provide opportunities for social
j
jplay. Mass media can also provide an occasion for social
i
play. When a family looks forward to a particular
television show that they watch together, or when a crowd
of friends goes to a movie together, or when a radio
»
I '
provides the music for a party, the mass media are invited
|to provide the background or the very foundation for the
jsocial gathering.
The impact: of symbolic interactionist sociology on thej
: . i
development of these concepts argues for simultaneous j
recognition of the distinction between self and society andl
i
the necessary mutual construction of both through one j
; another. Cooley's (1902) argument that self-concepts are '
developed as the result of social interaction, for
i
instance, suggests that self understanding cannot be j
i
1 «
jacquired without simultaneously acquiring or employing |
|one's understanding of others. One's orientation towards j
| i
social interaction, as well, would then suggest some 1
understanding of society, and by implication some
understanding of one's self as an actor in society (Mead,
1934).
i
I
At the individual level of analysis, Ball-Rokeach
(1985; Ball-Rokeach et al.. 1984) proposes that the pursuit
of goals that can be broadly categorized in the dimensions
t
jdefined above can result in relations of dependency between
jpeople and the mass media. These dependency relations
jdevelop when people perceive the media to control resources’
j
( that are central to the satisfaction of these goals.
Dependency relations develop without the individual
necessarily perceiving their development, as with many
L
jsocial relations (e.g. relations of production). When
jasked, people may be able to recognize the importance of
jthis or that media resource in their daily lives, but this
17j
importance may not be particularly salient to them day to |
I
day. i
I
I
In a given society, the antecedents of an individual j
media system dependency relation include the structure of
I
J
}the media system and the media system's structural
I
•dependency relations, the social environment (particularly
the amount of ambiguity and threat in that environment), j
media activity (i.e. the particular agenda or interest of ;
i
the media at a given time), the access one has to various
j
imedia, and the goals one has and "the extent to which these
t
goals are contingent upon the resources of the media
system" (Ball-Rokeach, 1985, p. 495).
j The outcomes of a media system dependency relation
I
i
jinclude increased attention to media content that addresses
| itself to the one's substantive motivations (Ball-Rokeach
I
j et al.. 1984; Grant, Guthrie, and Ball-Rokeach, 1989;
I
jcolman, 1990), and increased susceptibility to changes in
ibelief and behavior after attending to media content (Ball-
Rokeach et al.. 1984; Colman, 1990).
i Media System Dependency Research
I
Research into the nature of media system dependency
relations began with the quasi-experiment reported in The
Great American Values Test by Ball-Rokeach et al. (1984).
iThe researchers produced an original thirty minute
I
television program for broadcast over commercial, network-
f
affiliated television stations in the Tri-Cities area of
eastern Washington state. The program, which featured
well-known television personalities Ed Asner and Sandy
Hill, provided viewers with information about the value
i
priorities of Americans. Included in the information were
statements designed to provoke a state of self-
confrontation in the viewers. Self-confrontation is the
i
jmethod developed by Rokeach (1973) to allow people
privately to examine their own value priorities in light of
information they have received about the priorities of
others or about the possible contradictions in certain
Ipriorities.
i
Media system dependency theory was used to develop
hypotheses about the audience members who were most likely
i
to (1) watch all of or part of the program, (2) be
Icognitively and affectively aroused during viewing, (3)
i
lundergo some change in values as a result of viewing, and
I
! (4) exhibit behavioral changes in the weeks following the
I
broadcast. Prior to the broadcast, the program was
;advertised and promoted in the television listings in a
manner that emphasized the potential for obtaining self and
social understanding through viewing. A scale to measure
the intensity of dependency relations was devised, with
i
! special attention to measuring the understanding
<
idimensions, and a sample of survey participants was drawn
|in Experimental City and in Control City, where the program;
t 1
|was not broadcast. j
I |
The researchers did not ask anyone to view the 1
i
program. The decision to watch or not was, presumably,
I
made by the residents in Experimental City according to the
same criteria they use to decide on any new or special j
'program. The viewing situation was "natural," in that
i
viewers were subject to the same random and ritualistic |
interruptions during viewing that they are always subject
!
to, e.g. crying children, meal times, telephone calls, and
conversation. Following the program, the people who made
up the previously surveyed sample were contacted by
telephone and asked if they had watched and if they had
been interrupted while viewing.
i
Ball-Rokeach et al. (1984) demonstrated that those
viewers whose survey responses indicated a more intense
i
understanding dependency relation with television were more
ilikely to tune in, more likely to undergo value change, and
more likely to contribute money and political support for a
jvariety of causes that were related to the values that had
i
jbeen targeted for change in the program (for instance, a
i
|"ban the bottle" ecology bill was associated with the
!
priority given the value "a world of beauty").
j
I The Great American Values Test supported some
J
essential hypotheses about the impact of dependency
20
I
relations on the experience of watching television and the j
potential for media effects when dependency relations are
particularly intense. Most important for present purposes,
the relationship between viewing practices and dependency
i
relations was established at the level of a specific j
program by the observation of significantly different i
levels of dependency intensity between viewers and '
I nonviewers of the program. If this relationship can be
generalized to day to day viewing, one may conclude that
television use is to some significant extent related to
dependency relations, and that the theoretical expectations
one has about dependency relations and television viewing
may deserve further analysis.
! One subsequent test of the television dependency
: relation involved regular viewers of a home shopping
j
! network (Grant et al.. 1989). With a scale to measure the
intensity of dependency across the full scope of dependency
i
irelations in understanding, orientation, and play, Grant et
' al. established relations between the intensity of
i
;dependency and viewing, purchasing, and parasocial
interaction. These results, in light of the previous study
by Ball-Rokeach et al. (1984), further support the link
between day-to-day viewing habits (in this case related to
a particular program type out of the control of the
researchers), dependency relations, and behavior that can
Jbe traced directly to the viewing experience (in this case '
purchasing).
I
A third investigation of media system dependency
relations and television viewing is reported by Colman j
i
j(1990). Like The Great American Values Test. Colman's !
j t
'study examined the effects of a television broadcast on j
l j
people allowed to view (or not view) the program under i
natural conditions—in their own homes and without intrusion
[from the researcher. The program, a made-for-TV movie
i
jcalled "Leap of Faith,” concerned a woman's struggle
against cancer through the use of holistic medicine,
jcolman outlined the effects process that Ball-Rokeach
| i
j(1988) had proposed, in which dependency relations may be
[activated by the content of a television program (or
[magazine article or top-40 song) that heightens arousal,
^which then increases the information processing after
!
exposure in which a person is likely to engage, all of
t
which make changes in belief and behavior more likely.
Colman (1990) examined the changes in the attitudes
and beliefs regarding the efficacy of holistic medicine her
'respondents reported following exposure to "Leap of Faith."
jshe found that "even watching just a small part of this
movie . . . led to stronger belief in the ideology of
I
holism" (Colman, 1990, p. VI.12). This effect was observed
despite explicit attempts by the movie's producers to
disclaim any advocacy of holism, and despite the dramatic,
[entertainment-oriented nature of the script (the movie was
I
not designed as a persuasive vehicle). Using a measurement
scale based on that used by Grant et al. (1989),2 Colman
found that the intensity of television dependency relations
(across a broad substantive scope) was the best predictor
of arousal to the content of the movie. Arousal, in turn,
predicted higher levels of post-exposure information
'processing (i.e. thinking about and talking about the
movie's subject), which then predicted the observed
increase in holistic beliefs.
The relationship between the intensity of dependency
relations and newspaper reading was probed by Loges (1990).
Loges made use of data collected by Ball-Rokeach et al.
I
: during their research for The Great American Values Test
|
'concerning the newspaper habits and dependency relations of
those responding to the surveys in both Experimental and
[control Cities. Loges demonstrated that dependency
relations account for nearly 50% more variance in
readership than the combined demographic profile variables
i
(traditionally studied in readership research. The
'understanding dimension of dependency was strongly related
I
1 __________________________
i ?
*The development of a single scale whxch would
ireliably measure the intensity of all six dimensions of
Imedia system dependency had been continuing since The Great
[ American Values Test. This issue receives more attention
[ in Chapter 4.
23,
i
to readership. Further, the terminal value priorities
I
(Rokeach, 1973? Ball-Rokeach et al.. 1984) of readers were
related to newspaper dependency relations such that social
values were ranked significantly higher by those with the j
t
most intense social understanding dependency relations.
{These results, in addition to providing more insight into ,
the profile of newspaper readers, support several of Ball-
i
Rokeach*s propositions regarding media system dependency
relations (1985) in a medium other than television.
■ The interplay between media system dependency
1 j
Irelations and other sources of information was probed by
Aydin, Ball-Rokeach, and Reardon (1990). In a study of the
informational resources used by breast cancer patients,
Aydin et al. demonstrated a high correlation between the
intensity of the self-understanding dependency relation and
reliance upon mass media or interpersonal resources for
social comparison. For example, for women in a patient
support group intense self-understanding dependency
{relations with family and friends was correlated with
i
i (
{informational dependency relations with family and friends
for purposes of social comparison. Similarly, in the same
group intense self-understanding dependency relations with
{television and radio were correlated with dependency upon
[those media for social comparison. The correlations
i
{between media self-understanding dependency relations and
personal social comparison resources, and between personal
self-understanding dependency relations and media social I
comparison resources, were not significant. The self-
! evaluations of the patients were shown to be the result of j
j ;
an interaction between media dependency relations, personalj
dependency relations, and the perception of information
gained from the media as ’ ’uplifting" or "depressing."
The intensity of broadcast media dependency relations
in the case of general pervasive ambiguity and increased
I
; environmental threat was examined by Hirschburg, Dillman,
iand Ball-Rokeach (1986). Pervasive ambiguity is the lack i
of sufficient information to define a situation (Ball- !
jRokeach, 1973). Following the eruption of Mt. St. Helens
!
I in the State of Washington, a survey was conducted in which
! respondents reported their information-seeking activity in
I
jthe days following the eruption. Use of broadcast media
| (radio and television) and use of interpersonal resources
i
(including relatives and friends contacted by telephone)
was compared. It was found that the broadcast media were
( overwhelmingly the most utilized sources of information by
. respondents from all social classes. This was true on the
|day of the event (May 18, 1980) and on the days following '
the event. The authors note that the normal variation in
media use among diverse social categories disappeared in
;the wake of the eruption. This observation supports the
'proposition that the media system is known to be central to
|the informational needs of American society as a whole, and
Ithat when the need for information becomes acute—as it does
! I
I in the instance of pervasive ambiguity (Ball-Rokeach, 1973)i
i
—members of society turn to mass media for the information '
I
that will help them make sense of their situation. j
Information Seeking and Threat !
; The media system dependency relation is not a "media
jeffect." Ball-Rokeach (1985) does note that any
i
hypothesized effects are more likely in conjunction with
intense dependency relations (this hypothetical process is
outlined in Ball-Rokeach et al.. 1984, and in Colman,
j1990). The study of MSD relations has focused on the
I
circumstances under which information from the mass media
may become more central to the goals of people, therefore
i
increasing the intensity of dependency relations. Ball-
Rokeach has proposed that one such circumstance is a
[
(condition where a person perceives that he or she is
i
|threatened by the environment—both the "natural"
jenvironment and the social environment (Ball-Rokeach et
i
al.. 1984). This proposition is similar to the
i
jexpectations of effects theories of linear relationships—
.positive or negative, depending on the theory—between a
i
jperson's perception of threat and his or her use of the
media.
It is important to note, however, that Ball-Rokeach
also associates an increase in the intensity of media
jsystem dependency relations with an increase in actual
’threats in the environment. In the following chapter I
iwill have more to say about the difference between the
{perception of threat and a hazard's actual potential, but
for now it is important to distinguish between the two.
According to Ball-Rokeach (1985), the amount of actual
threat has an impact on the activity of the media—e.g. what;
gets on the agenda. Media activity is one antecedent of j
individual level dependency relations. But the presence of1
a hazard in the environment is no guarantee that a person
will experience threat, for the hazard may not be perceived'
iat all, or a person may not feel vulnerable to it. The
proposition that the amount of actual threat in the
environment will lead to more intense media system
dependency relations is distinct from the proposition that,
at the personal level, the more one perceives threat in the’
i
; environment the more intense his or her dependency
relations will be.
I When Ball-Rokeach et al. (1984) proposed that
! perceptions of threat in the environment will intensify
i
media system dependency relations, they assumed that the
i
jperception of threat would naturally lead to the goal of
understanding the threat and acting in a way to minimize
27!
i
it. This goal, the acquisition of information regarding j
!
the threat and regarding appropriate action in response to j
i ' ;
|the threat, may seem like a rational reaction to perceived i
Ithreat. In fact there are other plausible alternatives. I
1 i
ilf one is truly frightened by the perceived severity of the!
i ;
(threat, one may not wish to hear any more about it. If one
jis afraid and convinced of the accuracy of his or her
[perception, additional information may be merely an
i
unwelcome reminder of the impending harm (Janis, 1958). On
i
the other hand, if one perceives a threat such as a weak i
economy or a crime wave, one may prefer to believe that
I
social institutions are addressing the problem and that
i
[there is nothing to be done at an individual level, thereby
■ i
I
avoiding responsibility for the threat. In these cases,
i
(the availability of more information is not as important as
i
[the goal of comforting oneself by avoiding that
i
I
information. Acquiring such comfort may be an action
orientation or self understanding goal that is not
conducive to media system dependency relations.
Beniger (1986) correctly observes that the capability
to process information is essential to control, and that
control is essential to life. Turning off a radio is as
much an attempt to process and regulate information as
turning it on; rejecting a magazine subscription as much an
attempt to control information as accepting the magazine.
f
1
28
J
To observe that, in response to threat perceptions, people .
may turn away from certain information is not to suggest j
that they are foregoing the opportunity to process
information or exercise control over their environments. |
Intuitively and empirically, we know that perception I
|of threat is not just a function of media relations. Media!
| >
relations most likely interact with other behavior that an 1
i
individual engages in, and with other social relations in j
which an individual is involved, to produce his or her
i
perception. Likewise, we know that the environment itself,
regardless of the individual’s behavior, contains inherent
threats that may be perceived but unaffected by media
i
habxts. With so many variables contributing to perceptions
!
;of threat, it may be that the best model of the
relationship between perceived threat and media system
dependency relations is not a linear model, but some other
sort. But what sort?
Herbert Simon introduced the concept of "satisficing"
'to account for information-seeking behavior (Simon, 1955).
Simply put, Simon proposed that a person ceases to seek for
additional information when the cost of that information
exceeds the benefits it is expected to provide. This means
that even when one knows that there is more information
available, one will not seek to obtain additional
information unless one perceives its utility to be of
greater or equal value to its cost. In mass media, an j
example could be the decision to install cable television i
I
i
in one's home. Installation is not a knee-jerk reaction to
the availability of the programming on cable, but instead j
the result of a cost-benefit calculation. The result, when|
i ■ ■
;graphed, is a parabolic, or bell-shaped curve where cost is
I
measured on the horizontal axis and information obtained on
I
jthe vertical axis. The form of this relationship is due to
|the factors that impact cost, such as one's available
i
resources, time, and the amount of information available.
The immediate transaction wherein money or some other
resource is exchanged for information focuses one's
f
attention on the material economic aspects of this
activity. There are other "costs" to consider, however.
jFor most people, there exists a point beyond which
information about a certain subject—even "cheap"
information—exacts a psychological cost and thereby
discourages further information-seeking despite what might
be an apparent need for more information. Information
seeking may also be curtailed when the prejudices of a
iperson are challenged by the information they begin to
■receive (Frenkel-Brunswik, 1950).3
| 3"There seems to be a general tendency on the part of
[unprejudiced people] to expose themselves to broad
experience . . . even at the risk of having to modify one's
(preconceived notions and of having to sustain conflicts"
(Frenkel-Brunswik, 1950, p. 464).
30
I
Put another way, the "rational man" who seeks
information in Simon's (1955) model may be brought to more
I
human dimensions by adding the boundary on rationality that
the perception of threat can represent. Simon demonstrated
i !
i
that "bounded rationality" was not irrationality, simply j
rationality that was based on imperfect information and \
limited resources. It is possible, with such a theory, to j
i
|conceptualize the boundaries and conduct research into j
jtheir properties and effects on behavior (Simon, 1978).
jWhen the bounds on rational information seeking are
extended to include psychological states, it may be
possible to similarly explore the impact of those limits,
j Organizational scholars have investigated such a
[problem. Daft, Sormunen, and Parks (1988) examined the
i
I
! relationship between environmental uncertainty (a concept
i
I that does not necessarily equate to threat, but which, for
| i
|a business, has distinctly threatening implications), the
I
|strategic value of information (a concept more directly
[related to the ability to reduce threat by gathering
jinformation), and environmental scanning (i.e. information
\
jgathering) by executives. They found that scanning
! increased as uncertainty about strategic resources
! '
[increased. Boyd (1989) elaborated on the findings of Daft
, et al. (1988) to demonstrate that executives who had a low
personal tolerance for ambiguity would engage in somewhat
I
L ___
less environmental scanning regardless of how uncertain !
|
their environment was, even regarding strategic resources, j
I
In short, patterns of information gathering in order
I
to reduce threat in the organizational environment are both
i
i
a function of the nature of the environment and of the j
'individual. Simon (1978) had proposed that where certain j
| i
Ihuman attributes such as "attention" and "mind" were
scarce, "information may be an expensive luxury, for it may;
!turn our attention from what is important to what is
unimportant" (Simon, 1978, p. 13). Actually, our attention
‘is turned to or away from what we recognize as important,
! whether the object is "really" important or not. Boyd's
i
j(1989) findings support a bell shaped curve where the
jvertical axis indicates scanning activity and the
jhorizontal axis is perception of uncertainty:
j [A] perceived lack of information, by itself, has
I no appreciable effect on scanning activity. When
combined with strategic importance, however, this
I variable has a significant effect on scanning:
! Paradoxically, a perceived lack of information
1 will lead to a decline in scanning activity.
(Boyd, 1989, p. 81; emphasis in original)
The decline of overall scanning activity at the apex
jof the curve is attributed, among other things, to a
jportion of the sample's inability to tolerate ambiguity
beyond a certain point, leading to reduced scanning
activity despite—or even because of—an ever more uncertain
environment. The apparent importance of more information
is, for some, not enough reason to contend with the
ambiguity to which a search for information exposes one.
I I propose a similar model of the relationship between
media relations and perceptions of threat. Furthermore I
propose that, while frequency of media use can be a helpful
concept, the media system dependency relation is a richer
conceptual definition of the individual's interaction with
jthe media. Ball-Rokeach (1973) has defined ambiguity as a
jcondition in which key information is absent. The lack of
I
|information in ambiguous situations can lead to stress, and
I
(therefore to behavior to reduce stress. Part of this
I
i
behavior is the acquisition of information. The work of
Simon (1955), Boyd (1990), and Janis (1958) provides
further insight into the information-related behavior of
people. I propose to examine the media relations that may
lead to behavior to see if patterns of media system
dependency relations are similar to patterns of information
{seeking described by Ball-Rokeach (1973), Simon (1955) and
|Boyd (1989). I would expect that media system dependency
I
i
|relations would be related to threat perceptions such that
moderate levels of threat perception are associated with
1
;the highest intensity and scope of dependency relations. I
believe this is due to the effects of various social
locations that place people in different positions from
which they may affect the environment and its threats
(Ball-Rokeach, 1985).4
In other words, media behavior and media system
dependency relations are related to perceptions of threat
in the environment, and all three are related to one's
ability to affect the threatening aspects of the
environment through one's control over financial,
I
!intellectual, and other resources. These relations combine
I
such that threat perceptions and dependency relations (both
intensity and scope), are related in the shape of
parabolas. Simon's (1955) concept of "satisficing" is
recalled by my hypotheses, as is Boyd's finding regarding
i
Ithe psychological limits of information processing ability.
I offer, by way of theoretical explanation, both a
psychological and a material basis for the hypothesized
behavior. Specification of these hypotheses will follow a
more precise conceptualization of threat in the following
I
[chapter.
4The concept of "social location" is used in this
dissertation as it is used by Ball-Rokeach (1985), i.e. to
! refer to the position in society one has as a result of
'one's age, sex, class, ethnicity, group affiliations,
Ireligion, and other characteristics which are often called
/'demographic." The variables by which social location is
most often identified I will refer to as "demographics,"
but when my emphasis is on the overall significance of a
particular combination of demographic variables, I will
refer to social location.
34!
Chapter Two: The Concept of Threat
I
Most of us believe we know a threat when we see one.
It is this confidence, perhaps, which has led researchers \
to forego a precise definition of threat even when their j
I
i
iwork directly involves this concept. After a review of j
literature related to threat, including research into !
i
stress, risk, and ambiguity, I concluded that threat has j
not been defined satisfactorily. I began to develop a i
| '
^conceptual definition which would capture the meanings
jimplied in earlier work involving threat and also express
the specific meaning I had in mind while conducting my own i
i
research.
{ I will review the shortcomings in the related
'literature in detail after presenting my own conceptual
definition of threat. For now, I will offer a brief
summary of those shortcomings. The stress literature seems
to avoid the problem of defining threat by not defining
this concept. While stress and threat are differentiated,
it is the concept of stress which is carefully
conceptualized and operationalized. The risk literature
;does not address itself to threats as such, but treats
i
hazards as probabilistic outcomes. The perception of those
hazards as threats is seen more as a barrier to the true
i
i
mission of risk analysis except inasmuch as that perception
i
affects the sort of risk communication needed to inform or
35
reassure the public. The ambiguity literature explores
some of the ways that situations become perceived as
threatening, particularly when information about the
situations is lacking. "Threat" is not a central concept
to ambiguity, however, and does not receive attention as
such. Despite their inability separately to provide a
satisfactory definition of threat, the three lines of
i
jresearch mentioned above do provide insight into the
requirements that a definition of threat must meet.
1
j I conceive threat to have three dimensions: danger,
jconjecture, and personal vulnerability. Danger. the
t
{exposure to loss or harm, is essential because there is a
t
[definitively negative aspect to threat— we are not
"threatened" with good health or good fortune. Coniecture,
a guess based on uncertain probabilities, is essential too.
The distinctly unpleasant nature of the danger in a threat
is that it is potential danger—a perceived condition
established by someone or something. We are threatened by
things we believe can or might come to pass. If one
(removes conjecture from a threat, leaving only information
J
:about danger to come (which, without uncertain
I
Jprobabilities, is now inevitable), one might consider the
[information a "warning" or an "alert", but "threat" does
i
[not seem appropriate for such a circumstance. We do not
t
tissue "tornado threats" or "hurricane threats" when we know
36'
I
'they are on their way. Finally, personal vulnerability is
I
necessary. One must perceive oneself to be personally
E
exposed to the danger if it comes to pass. Personal
f ,
exposure is not limited to physical harm to one's person. ;
i ;
When a family member is exposed to danger, the rest of the •
i ■
jfamily may be said to be threatened by the loss of that
iperson. Such a threat, felt by a family member who is not
limmediately in danger of physical harm, is nonetheless
!"real. "
I
[
| By this definition, danger, conjecture, and
vulnerability are all necessary and collectively sufficient
i
conditions. A dangerous and uncertain situation which
affects a person is not threatening to him or her if the
danger and his or her vulnerability are not perceived, and
i
'there is no grounds for conjecture about the outcome. One
i
jean believe that a certain eventuality is dangerous in
principle, and that the actual probability of the event is
juncertain, yet if one believes that he or she is not among
Ithe people who will suffer from the danger, one does not
feel threatened. For instance, someone in Los Angeles may
I
jfeel sympathetic, but would not claim to be threatened, by
a bomb threat to an airline in Germany or a government
building in Lebanon (unless, as noted above, one's family
jwere exposed to the distant danger). If one perceives only
jdanger and vulnerability but considers the harm he or she
37^
is exposed to inevitable, one would not claim to be
"threatened;" one is more likely to say he or she is
"endangered," as when one is diagnosed with a serious
illness. If one perceives an outcome to be uncertain and
oneself to be vulnerable to it, but does not consider the
situation dangerous, one is not "threatened" by that
situation. A lottery one has entered is an example of such!
a combination of conjecture and vulnerability.
The components of threat may combine in varying
degrees to determine the severity of the perceived threat.
A person may perceive more or less danger and greater or
i
I
less personal exposure to the danger (e.g. loss of one's
t
[home in a fire compared to loss of one's child), and may be
^able to make a guess as to the outcome with greater or less
I
i
! certainty. Danger, conjecture, and vulnerability are taken
into account when people evaluate a threat as severe or
i
mild.
Furthermore, threats are conceived as personal or
social, differentiated according to the levels of society—
i
Ifrom the person to the whole society—which they put at
jrisk. A direct, personal threat (e.g. an eviction notice)
; requires a different sort of coping strategy than a
jgeneralized, social threat (e.g. an economic recession).
Personal threats often imply a specific action which must
be taken in order to alleviate the danger. Social threats,
I
|or threats posed by nature (e.g. volcanoes or hurricanes)
rarely can be assuaged by the actions of any particular
{individual. Further, some threats are acute, and once they
I
jare addressed they go away. For example, when a child is
[infected with chicken pox, generally the threat of that
disease is removed for the future. Other threats are
chronic, however, and can only be alleviated through
continuous effort. The threat of an economic downturn is
always present to one degree or another, and whole
[financial institutions exist to monitor the economy and
Jaddress that threat daily.
! I am concerned with chronic threats that may be
perceived at the personal or social level. Such threats
jinclude economic crises, injury or illness, natural
disasters, and environmental problems. I am interested in
I
knowing how perceptions of such chronic threats are
t
correlated with the information-seeking behavior and
dependency relations people engage in with the mass media.
*
j The conceptualization offered above has roots in three
I
other strands of research: Studies of (1) stress, (2)
risk, and (3) ambiguity have made use of a concept of
i
threat which is similar to that defined above. A review of
;some pertinent literature, however, demonstrates that
threat is rarely conceptualized with the same care that the
I
i
terms numbered above receive.
Threat and Stress
In the 1950's, the study of stress and stress
t
management became a primary concern of psychologists and
■social psychologists. Janis (1958) attributed this
■interest to a variety of new stress-stimuli in American
i
jlife. For example, the fear of Communism, nuclear weapons,
j
jtechnological change, and a rapidly expanding, complex
'economy are some of the environmental factors which Janis
jsuggests may promote stress. Each of these environmental
Ifactors can be perceived as threatening, and indeed a
concept of threat was needed to account for the stressful
'impact of these factors.
| [W]e distinguish a class of stimuli which are
more likely to produce disturbance in most
i individuals. The term stress has been applied to
i this class of conditions . . . which by their
| explicit threat to vital functioning and their
i intensity are likely to overload the capacity of
] most organisms' coping mechanisms (Basowitz,
; Persky, Korchin, and Grinker, 1955, p. 7,
| emphases in original).
|
I [The] major phases of psychological stress . . .
I [include the] threat phase, during which the
person perceives signs of oncoming danger and/or
receives communications of warning which are
likely to arouse anticipatory fear . . . (Janis,
1958, p. 7, emphases in original).
It seems apparent from the examples above that threat
was conceptualized as an element of stress and that threat
jis distinguishable by its informational quality—a threat is
a "meaning" or a "communication of" something, not the
thing itself (c.f. Withey, 1956).
Janis (1958) studied the reactions people exhibited to
the stress caused by impending surgery. He found that the j
threat represented by surgery-whether major surgery or j
minor—was dealt with in a variety of ways, largely j
| i
^represented by the information-gathering processes patients J
I !
engaged in. He found that patients with a great amount of |
fear tended to ask few questions and to predict the most !
I
|dire of consequences. They were uncooperative with and \
I I
’ suspicious of the medical staff, and likely to believe that:
j j
the staff were deliberately misleading patients regarding
the seriousness of the threat of surgery. The patients
l
t
with few fears also asked few questions. They believed
I
that all would be well, and tended to anticipate with
relish the post-surgical care and attention they would 1
receive. They were pleasant to the staff and undemanding
as they awaited surgery.
The moderately frightened group were unlike the others
in their pre-surgery behavior.
Instead of dismissing the impending operation as a
trivial or joking matter, they may be inclined to seek
information about the threat and to think in terms of
mitigating factors (Janis, 1958, p. 306, emphases in i
1 original).
I
j Those in the moderate fear group were the most likely
jto believe that (1) they did not know all there was to know
jabout the upcoming operation and that (2) knowledge could
•help them prepare for the event and its aftermath. In
4i!
i
fact, the threat that surgery represents is not so much the;
i
procedure itself but the post-surgery period (Janis, 1958).(
It is the recovery period that is the subject of ;
conjecture, even if the operation itself is a success on
medical grounds.
By the mid-1960's, research which drew on Janis' I
studies of stress attempted to clarify the relationship j
between threat and stress. In a series of laboratory
experiments, Lazarus et al. (Speisman, Lazarus, Mordkoff,
and Davison, 1964; Lazarus and Alfert, 1964; Lazarus,
i
Opton, Nomikos, and Rankin, 1965) referred to the process
of "short-circuiting” threat by preparing the audience to
view the content of a film in various distinct ways, a
process they referred to as changing the orientation of
i
jsubjects towards a stress stimulus (a film depicting severe
|injury). In fact, no measure of threat is ever presented
!in these studies, and the dependent variable is invariably
I
jstress (as measured by physiological reactions such as
I
jgalvanized skin response and heart rate). In each study,
the stimulus is presumed to be among those Basowitz et al.
(1955) consider to be reasonably universally threatening.
Lazarus et al. conclude from their series of
experiments that it is possible, by manipulating the
orientation of people towards stimuli which are normally
perceived as threats, to reduce the stress which results
I
(from exposure to these stimuli. Their manipulations
involved informing the audience either during the film
j(e.g. through voice-over, Lazarus and Alfert, 1964), or
before the film (e.g. in a live introduction, Lazarus et
i
al.. 1965) that the film was to be viewed as a scientific
display (encouraging intellectual distance) or as an
^interesting but harmless event (encouraging denial of the
evident brutality). In present terms, this may be seen as
j
manipulation of the danger and vulnerability aspects of
I .
threat.1
Lazarus et al. concluded also that the perception of a
stimulus as threatening depends upon the information one
I
has about the stimulus and the orientation one adopts
l
towards the stimulus based on that information. In its
extreme, this line of reasoning can counter the assumption
made by Basowitz et al. (1955) that there are explicit and
i
| 1In the cases of these films, the audience's estimation
Jof the probability of the graphic violence actually occurring
to the viewer can certainly be presumed to be low. The films
|dealt with a Polynesian ritual mutilation (Lazarus and
Alfert, 1964) and with industrial accidents in a sawmill
(Lazarus et al.. 1965). The subjects were college students
((Lazarus and Alfert, 1964; Lazarus et al.. 1965) and airline
executives (Lazarus et al.. 1965). There is probably a
vicarious nature to all cinematic representations such that
the audience feels a level of involvement which suspends
normal conjectural calculations. Fear is heightened by the
temporary inability of the audience to conclude with
certainty that they are not, in fact, vulnerable. Still, it
appears that there was no explicit attempt to manipulate the
audience's conjecture beyond the inherent confusion of a
Cinematic experience.
universal threatening stimuli that contain inherent danger j
to vital personal functions, since changing one's
j
orientation may be sufficient to reduce the perception of j
i
i
(threat. Another investigation of stress may shed some
light on the interaction between orientation and threat
evaluation. |
i
Landau (1989) surveyed Israeli Jews to determine their!
^'subjective perception of well-being" in the face of
["objective social stress factors" including a fragile
[economy, continuous national security concerns, and
I
(immigration tensions. The premise of this study is that
there exist social conditions that are objective stressors,!
that are presumed to be threatening to well-being. A
i
|society such as Israel's, faced with constant and diverse
|threats, cannot simply reorient its appraisal of the
I
f
jenvironment through different information and subjectively
reduce the threat. Landau (1989), in fact, demonstrated
f
the pervasive ability of the objective social world and its
attendant concerns to shape subjective perceptions of
I
well-being. In their study of Israeli media habits, Katz
i :
et al.. (1973) remarked upon the rather constant attention
Israelis paid to news briefs due to their conflicts with
the Arab nations. While the news reports may, on most
i
joccasions, reassure the Israelis that they are not
Ipresently in a shooting war, the receipt of this
(information does not seem to change the overall perception
jof the Israeli that his or her country is threatened. This
I
is due, Landau concludes, to the constant presence of
jobjective stress factors at work in Israeli society.
j Bandura (1977) has demonstrated the ability of people
1
,to overcome phobic reactions to stimuli by "learning” to
suppress their fear. Social learning-modeling one's
I
behavior after the observed behavior of others—allows
i
people suffering from irrational fears to overcome the
i
worst effects of these fears by observing the behavior a
non-phobic might engage in. For example, people with
chronic fear of snakes might watch a film of a snake
handler surrounded by snakes. While at first the film may
,frighten the phobic, repeated viewing may reduce the overt
j
demonstration of fear and reduce the incapacitating
reactions the phobic has demonstrated in the past. The
fear of snakes may remain high, but the phobic's method of
dealing with that fear may be changed to a less
pathological mode (Bandura, 1977). The repeated exposure
to a non-threatening stimulus may reduce the phobic's usual
presumption of threat when the stimulus is encountered
naturally. In present terms, all three dimensions of
threat may be reduced through observational learning. The
danger presumed to be inherent in a phobic's fear object
may be reduced (though perhaps not eliminated), the
certainty of personal harm may be reduced, and the sense of
personal vulnerability may be reduced as well.
Two important conclusions regarding information and
threat are suggested by the literature on stress reduction:
(1) There is a relation between sources of information and
perceptions of threat, and (2) People in threatening
'conditions which they perceive to be of varying levels of
I
severity seek different information according to various
criteria.
I It appears that the perception of threat is largely
I
jaffected by the quantity, quality, and context of the
I
(information one receives. Whether from interpersonal
|
jsources (Janis, 1958), institutional or authoritative
sources (Janis, 1958; Lazarus and Alfert, 1964; Lazarus et
al.. 1965), or from mass media (Landau, 1989), one's
relation to the source of information is an important
variable in determining the way the information is
I
[understood and the affect the information may have on one's
j
Iperception of threat.
Depending on how severe the threat is perceived to be,
jand how vulnerable a person feels, he or she may tend to
prefer information about different aspects of the threat—or
prefer no information at all (Janis, 1958; Lazarus et al..
1965; Bandura, 1977). If the threat is perceived to be
relatively acute, and particularly dangerous, the
information one seeks may be oriented towards reassurance
jthat after the disaster he or she will be missed (e.g.
jdeath fantasies) or treated with sympathy (Janis, 1958).
The threat is not reduced by such information; it is taken
for granted as plans are made to deal with the aftermath.
Chronic threats of high perceived severity may encourage a
•pattern of information seeking which is devoted to constant
i
jreassurance, not to elimination of the threat (Katz et al..
;1973) .
In studies of mass media, the relations between the
source of information and the goals of the information-
seeker are addressed by media system dependency theory.
i
[Ball-Rokeach (Ball-Rokeach et al.. 1984; Ball-Rokeach,
i
jl985) has proposed that media system dependency relations
l
jresult from the recognition by a person that a medium
controls information resources which are perceived by the
person to be crucial to the satisfaction of personal goals.
If one's specific information goals are related to threat
perceptions, it follows that perceptions of threat are
related to the media system dependency relations in which
I
people are engaged. Further elaboration of the
lexpectations one may have regarding perceptions of threat
j
and media system dependency relations will follow a review
of literature treating threat as a concept related to risk.
Threat and Risk
Information about an impending danger can be perceived
jas threatening under the circumstances of conjecture and
vulnerability already described. The same danger can be
perceived as a "risk," however, (1) when perception of the
{danger is contextualized by hopes for positive as opposed
jto negative outcomes, or (2) when attempts are made to
!
manage the conjectural dimension by reducing the
uncertainty to known (i.e. not uncertain) probabilities
j(Cancian, 1979; Douglas, 1985).
j While the context of gain and loss is occasionally
japparent in risk literature, it is the context of
management which predominates. As Douglas observes, "a
great deal of risk analysis is concerned with trying to
■turn uncertainties into probabilities" (Douglas, 1985, p.
42). Introducing a conference on risk communication at the
i
[Annenberg School for Communication in 1988, Rogers offered
I
,this definition: "Risk is the likelihood of a hazard
causing harm to an individual and/or what they value"
(Rogers, 1988, p. 10, emphasis in original). The focus of
this definition on "likelihood" is important, and typical
!of the literature (where definitions of risk itself are
rare, but where attention is often focused on the
I
Calculation of likelihoods). Douglas quotes the World
J
I
Health Organization:
48 I
Risk is a statistical concept and has been j
defined by the preparatory committee of the
United Nations Conference on the Human
Environments as the expected frequency of
undesirable effects arising from exposure to a
pollutant (W.H.O., 1978, p. 19, cited in Douglas,
1985, p. 20). j
jThe more a threat to public order or health can be j
represented in terms of its probabilities, the more likely j
lit is to be characterized as a risk. MIn simple terms, it
i ;
lis 'risk' insofar as you know the probabilities of various
joutcomes, 'uncertainty' insofar as you cannot specify the
i
probabilities" (Cancian, 1979, p. 3; c.f. Knight, 1971).
< ’
s
jAmong the topics addressed in risk studies are AIDS
^exposure (Freimuth, Hammond, Edgar, and Monahan, 1990),
i
jvarious "disasters" and natural hazards such as
'earthquakes, tornados, floods, radon gas, and volcanic
( eruptions (Sandman, 1988; Rogers, 1988; Burton, 1978;
[Perry, 1988), and societal hazards such as crime, nuclear
t
power and automobile traffic (Short, 1984; Sandman, 1988).2
i
| The literature on risk perception and risk
communication makes the differences between risk and threat
t
clear. Douglas (1985), reviewing the literature on risk
perception, notes that people do not live day to day with a
constant cognizance of the hazards around them. Their
2The broadest definition of "risk" is that offered by
James Short; "[R]isk is the probability of some future
event" (Short, 1984, p. 711). At this level of
generalization, the concept begins to lose its usefulness.
49
l
perception of risks "can never be analyzed with an |
l
inventory of the physical features of events, the scale of ,
damage, suddenness, or duration" (Douglas, 1985, p. 28)
l
because these physical features are placed in contexts j
| which must be accounted for. One such context, of course, J
I !
I ' ■
I which plays a critical role in risk perception, is that of
| 1
potential gain. j
To illustrate the context of gain and loss, one may
jconsider casino gambling (Blascovich, Veach, and Ginsburg,
1
1
!1973). While a bet of $5 is likely to be lost (endangered)
i
i
based on probabilities calculated by a vulnerable gambler,
the basis of decision-making is consideration of potential
'gain. The threat of loss exists, but the threat is
perceived as "risk" because there is hope for an outcome
|which is a net gain for the risk-taker. Compare gambling
I ;
!to the threat of an earthquake and the difference is clear.
i
If an earthquake does not occur on any given day, those who
i
have not suffered do not experience a net gain due to that
non-event. If the threat of an earthquake is realized,
however, the only potential outcome is a net loss. The
gambling metaphor is often applied subtly in arguments
regarding ostensibly dangerous technologies, such as
i
!nuclear power or military equipment. Once a nuclear power
i
'plant is operational, the energy it produces every day is
jweighed against the threat an accident represents.
50
Protests during construction can emphasize the threat the J
plant represents (as during the debates over Seabrook in
New Hampshire, where the protests concentrated on
evacuation plans in the event of a disastrous accident), j
while the utility emphasizes the potential benefits of j
I
energy once construction is complete. These debates can be'
understood as attempts to shape public perception of
nuclear power as a threat or as a risk.3
Another distinction between risk and threat is the
i
context in which the potential for danger is placed. When ;
i
the potential for harm becomes calculable with some level
of precision, a threat may be perceived more as a risk.
This context of coping, or management, does not concentrate
on the potential for a net gain—although that fundamental
difference between threat and risk still applies—but
instead concentrates on the nature of the conjecture
involved. For instance, when the Federal Reserve Board
lowers the interest rate it charges banks, the decision is
generally reported as an attempt to avoid an economic
3Stephen Klaidman (1991) complains that
environmentalists too often resort to the language of crisis
1 to describe environmental hazards and not often enough to the
jlanguage of cost/benefit analysis or risk management.
I Commenting on a New York Times poll that showed that 80% of
'Americans agreed that the environment should be protected
!regardless of cost, Klaidman observes:
| Such sentiments, in a nation that already spends $90
1 billion annually on pollution control, cannot be the
product of a rational approach to environmental
problems. (Klaidman, 1991, p. 73)
L
51
]
recession by risking inflation (e.g. Risen and Johnson,
1990). Reports are accompanied by summaries of the
economic indices available to the Federal Reserve Board
I
which provide the bases for its decision. Such a context
i
provides reassurance that economic downturns are manageable
i
and, unlike earthquakes, are not arbitrary occurrences i
i
(i.e. are not threats as such). To most people, however,
I ;
|the economy is experienced in a less global context. Plant
!closures are threats to those in the working class because
I
their perceived ability to affect the probabilities of such|
i
, I
events is generally low. Decisions affecting the national;
i
economy are made by people whose structural position in
society allows them to view their decisions more as risks
than as threats (or responses to threats). i
I AIDS is considered in terms of "high risk groups" in
which the probability of exposure to the HIV virus is
I considered greater than in the general public. Education
about AIDS focuses on identification of these risk groups,
identification of methods to prevent transmission of the
i :
1 virus, and information regarding the symptoms and
4In the mid-1980's General Motors attempted to close its
manufacturing plant in Van Nuys, California. The workers of
the plant resisted this move with surprising vigor, even
offering to buy the plant and to manufacture automobiles
independently. Ultimately the plant remained open throughout
|the decade. In July, 1991, GM once again announced its
,intention to close the plant. The example of Van Nuys is
pointed to by labor activists as a unique example of the
ipower of collective action.
development of the disease. The first two are clearly
i
related to control of probabilities (Stiff, McCormack,
i t j
i Zook, Stem, and Henry, 1990).
Unlike AIDS, in which any exposure to the virus is
discouraged, much risk assessment is devoted to calculating
I
"acceptable" risks when exposure to chemicals or natural ;
I |
disasters (or natural hazards such as radon) are analyzed. ,
<
The calculation of acceptability is not possible only on
the basis of physical facts about probability. Society
introduces a set of variables which include morality,
!
justice, public order, and other such concerns (Douglas,
r
1985; Nelkin, 1988; Watzman, 1990). Scientific calculation
is clearly appropriate to the concept of risk, but
intuitively unappealing for threat. Who would acquiesce to
I an "acceptable level of threat?"
i
Douglas (1985; Douglas and Wildavsky, 1982) is
critical of most studies of risk perception for not
jaccounting for the broad context in which natural and
I social hazards are understood.
[
The accepted theory of risk perception maintains
that the rational principle of selection would
I combine the probability of an event with its
I value. However, people tend to fasten attention
i on the middle range of probabilities. Risks
; clamor for attention; probable dangers crowd from
! all sides, in every mouthful at every step. The
rational agent who attended to all of them would
be paralyzed (Douglas, 1985, p. 59).
53
1
The business of everyday life makes the sort of risk
analysis in which most specialists engage impossible for ,
the average person. People do not want to think about I
!
their lives as a series of gambles. This may lead to
"irrational" disregard of certain hazards (e.g. radon gas, (
Sandman, 1988), which upon closer examination is an attempt!
j
to avoid being forced to look upon everyday life as a j
"gain" and mere avoidance of death or disease as a victory.
Asked about radon, people can identify it as a danger, and .
i
recognize their vulnerability, but refuse to identify the '
risk to them as acutely high even when they live in an area
I
with a particularly high concentration of the gas. For
present purposes, it may be said that they recognize the
threat, but do not perceive radon as a risk they are
running in order to "gain" another day of life in their
neighborhood.
| I reserve the concept of risk for those phenomena
I
wherein (1) there exists a possibility of benefit in
exchange for exposure to possible harm, and (2) one
attempts to assess precisely the probability of loss in
i
order to rationalize decision-making. Threat is
distinguished from risk by both the absence in threat of
any implied potential benefit and the more conjectural
nature of the probabilities in threats. Colloquially, the
[
term "at risk" is an exception, in that this phrase is used
I
\
J
54
generically to refer to any danger, regardless of which of
1
our present conceptual definitions the danger fits most
I
closely.
The specific economic, health, environmental, and
!social threats presently studied are periodically the
subject of research as risks (e.g. "risk assessment" or j
"risk communication"). My classification is meant to
emphasize that I do not assume that my population conceives;
of these hazards in the rationalized manner of risk ■
i
assessment, but in the manner of more vaguely defined ’
events, filled with foreboding, to which people feel
<
vulnerable in varying degrees. I am not, in other words,
challenging the conclusions or assumptions of risk |
I :
assessment because I am not evaluating the "real-life"
.likelihood of any of these events coming to pass. While I
will not attempt to account for all the variance in the
i
jsample's estimates of vulnerability, I suggest that
|patterns of media system dependency relations and, perhaps,
1
media use, will be associated with perceived vulnerability
Ito chronic threats. Following a discussion of threat and
the concept of ambiguity, I will present specific
i
hypotheses regarding these relations.
i
j Threat and Ambiguity
| Ambiguity has been defined as the absence of a
i
"definition of the situation" (Ball-Rokeach, 1973). McHugh
(1968) identified emergence and relativity as the ;
j
components of the process of defining a social situation.
[
Drawing on symbolic interactionist theory (Mead, 1932, j
i
1934; Blumer, 1962) and on contemporary functionalist j
I (
'attempts to account for the development of shared meaning j
j I
I in society (Schutz, 1962; Merton, 1957), McHugh proposed a !
i j
description of the definitional process in which temporal
and spatial concepts interact to form a definition.
Emergence is a temporal concept (McHugh, 1968).
Emergence refers to those sense-making devices which
require a conception of time—past, present, and future.
For instance, "theme" is one property of emergence, and
!refers to the desire of a person to recognize a pattern in
interactions. This recognition depends upon the notion
1 F
that certain interactions have either been engaged in
! f
before, or will be engaged in again, and are therefore
! ;
subject to expectations or rules which serve to reduce the •
ambiguity which might otherwise surround the interaction
(Garfinkel, 1963).
Relativity is a spatial concept in McHugh's
definitional process. Relativity refers to the immediate
relations between those taking part in the interaction, and
the expectations one may have of an other based on
I >
knowledge of the other*s social position. "Typicality” is
i
ione property of relativity, and refers to the ability of a
56
person to determine whether the interaction is the sort of exchange
that is normal for an other of this or that type. Naturally, decisions
about typicality inevitably introduce a historical or temporal u-
dimension, which clouds the distinction between emergent and relative
I
concepts. McHugh (1968) acknowledges this, but proposes that an •tb. j
1
attempt to separate the two with too much vigor would be ultimately *
fruitless. The roles of emergence and relativity in the process of j
definition as described by McHugh (1968) are complementary, and while
his analysis depends on the ability to recognize one from the other,
he does not propose that they may operate completely independent of
one another.
To McHugh, failure to achieve a definition of a situation leads
to anomie, the normlessness which Durkheim considered the ultimate
i
i
| manifestation of social trauma at the individual level. It is the !
|
: prospect of anomie which is the clearest link between threat and '
i
1 . . . . . i
ambiguity as discussed by McHugh (1968). That is, ambiguity itself is :
f
: not necessarily threatening, but the absence of a definition of a
i
| situation can lead to distinctly unpleasant consequences under certain ;
1 conditions. McHugh was interested in describing the process by which >
people achieve definitions — leaving aside any concern with the (
|
"truth” or accuracy of those definitions. He posited that i
57:
i
i
the determination to achieve definition before continuing
in an interaction stems from the desire to avoid anomie.
!
| Ball-Rokeach (1973) conceptualizes ambiguity more |
i
thoroughly than McHugh by introducing temporal and spatial j
properties to ambiguity itself. Ball-Rokeach distinguishes
between "focused ambiguity" and "pervasive ambiguity."
i
"Focused ambiguity occurs when persons are unable to 1
I
jdetermine an appropriate strategy of action to deal with
known contingencies of a situation" (Ball-Rokeach, 1973, p.,
378). In a condition of focused ambiguity, concentration
seems to center on what McHugh would call spatial I
I
'properties or relativity—i.e what is typically done, what
is likely to happen, what are the causal relationships,
what is effective, what is proper, and what is consistent
jWith others' expectations within this situation.
f
An example of focused ambiguity is the situation of a :
i
baseball batter facing a curveball. When the ball leaves
the pitcher's hand, it appears to be aimed at the batter's
i
head. As the ball gets closer, the batter must decide
whether the pitch will in fact "break" and move over the
plate. If the batter begins to suspect that the ball will
not break he must duck or lunge out of the way (or get
hit). There is no ambiguity regarding what is going on, or
jeven why. There is, nonetheless, a considerable amount of
i
ambiguity as to what should be done about the situation.
58'
Contingencies—the rules of baseball—are known, but the best!
strategy is unclear.
i
As opposed to focused ambiguity, "persons experiencing!
pervasive ambiguity must resolve fundamental questions of j
meaning, such as what is happening and why" (Ball-Rokeach, j
1973, p. 379). In this condition, the temporal, emergent !
1
properties identified by McHugh seem more relevant. ^
jPervasive ambiguity calls upon people to consider what
I
theme broadly fits the situation, and then to observe the ;
elaboration of that theme through the course of the
> i
situation, and to test the theme's applicability through
i ;
the search for patterns of interaction which are familiar, j
McHugh suggests that the inability to satisfactorily
resolve the questions posed by the emergent dimension
forestalls any attempt to address the relativity dimension.
Similarly, Ball-Rokeach observes that in a condition of
pervasive ambiguity "persons have no direct way of knowing
I
i
!how they could or should interact with others or with the
I
environment, and thus the capacity to engage in meaningful
social action is impaired" (Ball-Rokeach, 1973, p. 379).
An example of pervasive amiguity, still using the
'baseball situation (roughly), would be if the batter had no
I
l
.idea what the rules of the game were or what his options
|Were as the ball raced towards him. Without a general
I theme in which to place this experience, and without any
prior experience to guide his expectations about what the
I
i
ball will do (let alone what the pitcher is up to, who the ;
catcher is, and what the umpire is for), the batter is !
!
unable even to say what is happening or (especially) why.
This befuddled batter's ideas as to the range of j
appropriate action to take may include assaulting the
catcher with the bat, since the catcher is an obvious
laccomplice of the pitcher (they wear the same uniform and
! i
have similar contraptions on their hands).
[ j
After conducting his experiments in social
interaction, McHugh concluded that it was, after all,
possible to address the spatial dimension in the absence of1
satisfactory temporal understandings. He concludes that in
practice the two dimensions are so integrally related that
it is possible for a person to probe one dimension, then
shift attention to the other when dissatisfied. Reading
jMcHugh's study in light of Ball-Rokeach's definitions of
|focused and pervasive ambiguity, however, it is possible to
conclude that McHugh's experiment establishes a situation
•which may appear to the subject as a problem of focused
ambiguity (testing a new method of psychotherapy) which, in
1
ithe course of the experiment, may be revealed to the
subject to be more pervasively ambiguous (when the
i
j"therapist'' responds randomly to specific yes/no
questions). It is consistent with the expectations of
60
Ball-Rokeach that, as McHugh's subjects are confronted with1
different types of ambiguity, they resort to different j
types of strategies to resolve their dilemma.
Ball-Rokeach's (1973) experimental condition placed
subjects in conditions of pervasive ambiguity immediately,
(with the aim of observing the strategies they employed to
I :
I adapt to the condition. After being told that they were to
jtake part in a laboratory experiment involving "directed
group activity," J
I
The subjects reported to a room, fifteen by j
twenty-five feet, which was void of all
| furnishings except for several travel posters on i
] the walls. . . . The room contained a one-way
mirror, a generous layer of dirt on the floor,
| and, unless a subject was the first to arrive,
other people who were strangers. None of the
personnel involved in the supervision of the
i experiment were in the room or in its vicinity.
! Thus, there was no way for the subjects to
I confirm their expectation that an experiment was
j in progress. Subjects were left in this
j situation for a minimum of thirty minutes.
(Ball-Rokeach, 1973, p. 381)
Confronted with the situation described above, Ball-
jRokeach expected that the subjects would grapple with the
;ambiguity by seeking information and by reducing tension.
!ln as much as the behaviors associated with both
‘ information-seeking and tension-reduction necessitated
social interaction, Ball-Rokeach hypothesized that subjects'
'would "select interaction partners on the basis of
(similarity," specifically age and sex (Ball-Rokeach, p.
i
;381). Finally, Ball-Rokeach expected that before ambiguity
6 If
I
could be resolved, subjects would have to arrive at an ,
"overt consensual acceptance of a proposed definition of
[the] situation" (Ball-Rokeach, 1973, p. 382), although as |
I
with McHugh the "correctness" of this definition was not j
; i
Ithe concern—only that the definition was supported by j
{others in the situation. Each of these hypotheses was
I
supported. I
The concept of ambiguity, particularly as studied by !
i
McHugh (1968) and Ball-Rokeach (1973), is related to threat
for two reasons. First, the conjectural dimension of
jthreat is clearly related to any concept of ambiguity or
uncertainty (conjecture is distinguished from these by its
i
I emphasis on a guess—i.e. a decision of sorts). Secondly,
i
and more importantly, the strategies both McHugh •
j ;
[(implicitly) and Ball-Rokeach (explicitly) describe for
coping with ambiguity are strategies of information
gathering. My present concerns, i.e. the use of mass media
[by people who consider themselves to be under various
j
[levels of threat, are clearly related to any process of
'information-gathering under conditions of ambiguity.
I *
Chronic threats may be related to pervasive ambiguity
I
!in that the temporal dimension is distinctly relevant to
these ongoing concerns. A broad definition of the
situation is not entirely lacking in the case of chronic
i
Ithreat (as it is typically in situations of pervasive
62!
i
iambiguity). On the other hand, it is impossible to focus j
on one ambiguous element in the environment, resolution of J
I
which will remove the threat. During chronic threats, as j
in conditions of pervasive ambiguity, people may choose to J
I j
tfind ways to gather information which will reduce the
uncertainty (and thereby reduce the threat by improving the
basis for conjecture, as in risk-assessment). It may be |
reasonable to expect as well, as in the case of pervasive
ambiguity studied by Ball-Rokeach (1973), that information
1
i
jseeking by those experiencing threat will be conducted on
the basis of perceived similarity.
Unlike conditions of pervasive ambiguity, chronic
threats are not without definition. The nature of the
j
|danger involved, and the evaluation of the person's
i
vulnerability, offer some definition to the situation and
allow for a discrimination of, say, an economic threat from
that of a natural disaster. It is here that I may
>
;reiterate that a threat is not an event that is occurring,
jbut one which is probable or portended. During an
i 1
earthquake, for example, there is often a period of
i
ipervasive ambiguity when those in the area try to ascertain
what exactly is happening. The explosion of Mt. St. Helens
in the State of Washington provided those living nearby
i
With their first experience with a volcano (Hirschburg et
i
l al.. 1986). Until the cause of the shaking ground and the
dark sky was located, pervasive ambiguity reigned. During |
(
the pervasive ambiguity, there may be discrete,
i
identifiable hazards such as falling rocks, ash, or j
collapsing buildings with which one must deal, but these !
are not "threats" in that there is no element of >
uncertainty about them— they are in fact falling. j
The threat of an earthquake is, in fact, not as j
jambiguous as the events accompanying an actual earthquake.
|There is little ambiguity regarding the danger an
.earthquake represents. When a person informs another that
i
.he or she feels threatened by the possibility of an
i
jearthquake, each person can reasonably recognize the sort
jOf danger that is under discussion. Still, there is enough
iambiguity about earthquakes that people may deal with their
Ithreat with strategies similar to those employed in
isituations of pervasive ambiguity. The conjectural aspect
lof the earthquake threat is largely centered around when
i
jthey may occur. Secondly, the question of where earthquakes
'will strike (which determines who is vulnerable), is
i
uncertain, although less so than the question of when. It
i
^s reasonable to expect that those people who feel under
(such threat to some degree day to day will search for
information about earthquakes and intermittently seek to
I
jrelieve the stress which the threat may inspire (Turner,
'Nigg, and Paz, 1986), both specific strategies utilized by
64
i
people under conditions of pervasive ambiguity
(Ball-Rokeach, 1973).
Summary !
In this chapter, I have defined threat as a perception
j
of danger and vulnerability, combined with conjecture about
the likelihood of the danger. I have reviewed the ways
I
threat has been used in studies of stress, risk, and
ambiguity with the aim of demonstrating that threat is a
concept distinct from those but related. Finally, I have
i
I
demonstrated the importance of information to the
perception of and reaction to threats. Information
l
gathering has been key whether threat has been studied as a
concept related to stress, risk, or ambiguity.
What about threat for its own sake? The next chapter
i
lelaborates the hypotheses and research questions which will
jdirect the analyses to follow. The support for these
i
[hypotheses will be drawn from the ideas in this and in the
i
[preceding chapter. I will attempt to draw connections
I
[between the media system dependency relations people have,
j
jtheir perceptions of threat in the social environment, and
Itheir location in the social structure.
Chapter Three: Research Questions and Hypotheses
The conceptualization of threat offered in the
preceding chapter provides an opportunity to examine, at
the most fundamental level, the relationships between
perceptions of threat in the social environment and other
social variables such as income and education, as well as
age, sex, and other demographic variables. While it may be^
possible to articulate a variety of hypothetical
' I
j I
'expectations regarding these relations and various :
| ‘
imultivariate combinations, I will offer a general research >
i
;question to summarize my interest in exploring these i
■relations:
I
I RQ1: What are the bivariate relationships between
; perceptions of threat and age, sex,
1 ethnicity, education, income, employment
status, and years of residence in one's
community?
i
Previous research into the nature of media system
dependency relations has sought to develop a scale which is
reliable across all six substantive motivations1 (Ball-
Rokeach et al.. 1984? Grant et al.. 1988; Colman, 1990;
jAydin et al.. 1990; Ball-Rokeach et al.. 1990). The
i
I research project which provided the data for this
idissertation included the development of a dependency scale
|(Ball-Rokeach and Grant, 1990; c.f. the following chapter,
f ^
| ’i.e. those of Social and Self Understanding, Action
land Interaction Orientation, and Social and Solitary Play.
and Appendix A). While the scale was pretested extensively;
as it was developed, it can not be simply assumed that the
empirical factors derived from the data collected will J
conform neatly to the theoretical dimensions of dependency J
relations. Analysis of dependency variables should follow 1
i
I
examination of responses to the scale to test the scale's '
'validity across the six theoretical dimensions:
I
f
RQ2: Does the dependency scale used provide a
reliable and valid means of measuring the
intensity of media system dependency
relations?
| The first chapter presented observations regarding the
f
manner in which information is sought, both from mass media
and from other sources, particularly under conditions of
threat. I observed that media theorists have often
proposed that perceptions of threat were associated with—
and perhaps caused—more intense media relations, either of
i
I use or of dependency. The second chapter explored the
I concept of threat in more detail, and reviewed (among
jothers) the research of Janis (1958) regarding stress and
i
I the threat of surgery. Chapter 2 also highlights the
I
difference between the amount of hazards an environment (
actually contains, and the perception of those hazards as
threats.
The search for information when one perceives threat
j
iseems to be bounded by certain conditions. These bounds—
(whether psychological (Janis, 1958; Boyd, 1989) or economic
67l
j I
j(Simon, 1955)—have been observed to create a distribution |
I i
j i
of behaviors related to information seeking which suggests ■
Ithe following hypotheses, contrary to previous conceptions i
in mass media theory: j
i !
HI: Perceptions of threat will be related to:
■ (a) the intensity of media system dependency I
j relations of social and self understanding, and
j (b) the intensity of media system dependency 1
relations of action and interaction orientation, !
; and to
I (c) the substantive scope of media system
; dependency relations, and
I (d) the media scope of media system dependency
I relations
1 such that lower and higher perceptions of threat
J are related to the least intensity and scope, and
that moderate perceptions of threat are related
i to the highest intensity and scope.
i
A partial explanation for the relations hypothesized
above lies in the social location a person occupies (Ball- ■
iRokeach, 1985). Those with better educations and higher
I
;incomes enjoy a broader access to information and other
|
|resources they may use to address the threats they face.
jThis expanded access may provide confidence that threats
j are less severe than they appear to be to the less educated
I
!or less affluent (and, in fact, certain threats are less
severe to the affluent, such as unemployment or rising
i
[medical costs). The danger of the various threats people
l
!face may be relatively equal across classes, but the
i
I
; vulnerability and the odds which inform conjecture are
different. While Janis (1958) noted no appreciable
;difference in the information-seeking of patients across
]class lines, his study focused only on behavior in a
I
jsurgical ward. The media behavior of his subjects prior to
|their hospitalization was not examined. In addition,
IJanis' subjects were all facing very distinct and similar
threats in very similar surroundings. Class would appear
to have been less of an issue in his study than in my
1
jbroader examination. The following hypotheses are
jsuggested:
j H2: Education and income will be negatively
correlated with perceptions of threat when
i dependency relations and demographic
variables are controlled.
The cushion provided by education and affluence can
[relieve much of the perception of threat, but these social
i
Jcharacteristics cannot entirely negate the impact of aging
I
on perceived vulnerability to threat. Much as the affluent
I
are in fact less vulnerable, the older one is the more
;vulnerable one becomes to social threats. For instance,
[health is a concern which increases with age, and cannot be
entirely addressed through affluence or education.
Vulnerability to crime increases too, since self-defense
becomes more difficult and one acquires more to protect,
whether it be family or possessions. The impact of age on
threat perceptions, moderated by other social
circumstances, provides the basis for the following
1 hypothesis:
69
H3: Age will be positively correlated with
perceptions of threat when dependency
relations and demographic variables are
controlled.
The relations between perceptions of threat and the
t
intensity and scope of dependency relations have been i
|hypothesized on the basis of expectations derived from the j
j i
jdefinitions of both variables and from prior research. !
Hypotheses la through Id treat threat perceptions as ,
independent variables which predict dependency relations. j
Hyptheses 2 and 3, on the other hand, propose to control ;
for dependency relations, treating dependency relations as t
independent variables predicting threat perceptions. I '
expect that the primary influence on threat perceptions
will be that of the demographic variables. Still, the
<
jintroduction in this study of the concept of threat
perceptions, and of a scale to evaluate threat perceptions,
.presents an opportunity to explore the relations between
|the variables under investigation from both ends, as it
were. It would be unwise to discard an opportunity to
check for reciprocal relations between dependency relations
and threat perceptions in the course of testing the second
and third hypotheses.
In the present study, no attempt is made to
I hypothesize the relationship between the amount of time
ispent using mass media and perceptions of threat, but it is
1 of interest to examine the relationship between these
variables. One obstacle to the development of hypotheses
i
I
regarding these variables is the sterility of the concept j
of use. The viewing environment, the attention level, the I
importance of the act of viewing, and the types of programs:
are all left unmeasured. Some of these contextual matters j
f
are addressed in a general sense in the dependency scale. j
Others, such as the viewing environment, are not 1
1
I necessarily relevant in this study.2 In short, while
there are not clear grounds for testing hypotheses
^regarding temporal use (i.e. the amount of time spent using
I
a medium, as opposed to functional measures of use), it is
i
of interest to examine the relationships between media
behavior and perceptions of threat, as well as between
media behavior and media system dependency relations.
! RQ3: What are the relationships between
J perceptions of threat and time spent in the
1 use of newspapers, magazines, radio, and
; television?
! RQ4: What are the relationships between media
| system dependency relations and time spent
j in the use of newspapers, magazines, radio,
and television?
I
' In the following chapter, I will describe the method
!
jused to collect the data in this study and I will describe
the methods used to analyze the data in order to test these
hypotheses and address these research questions.
! 2,,Effects studies" generally require such a measure to
account for the variance explained by distractions and
unforeseen interruptions.
71
Chapter Four: Methods
The Development of a Media System Dependency Scale
In September of 1988 I joined a group of researchers
and students led by Professor Sandra J. Ball-Rokeach at
the Annenberg School for Communication. The research
centered around a large-scale project to , , map, , the
dependency relations in two communities at the individual,
group, organization, and system levels. This ambitious
project was divided into specific tasks, and the chief
task for those of us addressing individual level
dependencies was to develop a single scale which could
reliably measure the intensity of dependency relations
across all six theoretical dimensions of dependency and
across a variety of media. By the fall of 1988, several
jstudies had been performed using increasingly more
reliable MSD scales (see Chapter 1 for a review of some of
these studies). Particularly, The Great American Values
Test (Ball-Rokeach et al.. 1984, which introduced the
first MSD scale), Grant et al.1s study of home shopping
(1989), Aydin et al.'s study of breast cancer patients
(1990), and Colman's study of the effects of a network
television broadcast on viewer's attitudes toward holistic
health (1990) each contributed knowledge about the way
items in the scales used were interpreted and how the
72
scale as a whole broke out into empirical factors when
subjected to statistical analyses.
Three considerations dominated our discussions, (1)
the statement which would serve as a general introduction
to each item, (2) the general appearance and format of the
scale as it called for responses regarding a variety of
mass media, and (3) the wording of the items themselves.
I
■Each aspect of our deliberations led to repeated tests of
prospective scales.
The key issue in the debate regarding the
introductory statement was the adjective which would be
used to describe the media. In The Great American Values
Test, respondents had been asked "how often do you use
television to . . ." followed by eleven items with a
response format of Often, Sometimes, and Never. Since
frequency of use was not seen to be the best
conceptualization of dependency, our attempts to revise
the scale centered around two competing introductions.
They were (1) "In YOUR Daily Life, How Helpful Are These
Media To . . ." and (2) "In YOUR Daily Life, How Useful
Are These Media To . . ." Our pretests showed a small
increase in reliability when "helpful" was the prompt, and
in the final instrument "helpful" was the adjective used.
Both terms are closer to the conceptualization of
dependency relations than a frequency of use measure, but
73
"helpful" adds a qualitative assessment of utility value
which is different from the more functional "useful."
This semantic debate, summarized here in a sentence or
two, occupied our attention for some time.
The appearance of the instrument was of concern as
well. We had the intention of measuring dependency
relations with at least four mass media—newspapers, radio,
magazines, and television—and with an eighteen item scale
we realized that the instrument could quickly become
tedious if each item were repeated four times. On the
other hand, the temptation for a respondent to fall into a
jresponse set if all four media were the subjects of one
prompt seemed a threat to validity. In the end, following
many pretests of the latter format, we decided to
construct the survey in the following manner:
In YOUR Daily Life, How NOT AT ALL EXTREMELY
Helpful Are These Media To: HELPFUL HELPFUL
| 1_____ 2______3______4_____5
Figure out what to buy?
Newspapers ................ 1 2 3 4 5
Radio ...................... 1 2 3 4 5
Magazines .................. 1 2 3 4 5
Television ................ 1 2 3 4 5
(The full instrument is found in Appendix A.)
During and between these debates, we concerned
ourselves with the wording of the items themselves.
Between September of 1988 and March of 1989, many items
were suggested and tests conducted of, first, a twelve-
item scale (c.f. Colman, 1990) and, then, an eighteen item
74
scale. In each case, the instrument was a pencil-and-
paper, Likert style measure. Test populations (besides
those described in published studies) included clerical
and maintenance employees at the University of Southern
California, employees of an insurance company in Los
Angeles, audience members of "The Tonight Show Starring
Johnny Carson" (while they waited in line to enter the
studio), members of the American Association of Retired
Persons, and undergraduate students at the University of
Southern California and at the University of Texas,
Austin. Focus group interviews were conducted with the
university employees and the insurance company employees.
These tests were conducted by members of the research team
using a succession of scales under development. As
results were analyzed the scale was revised and tested
again, and by February we had developed an eighteen item
scale we felt confident enough to use in a full pre-test
using random samples from Colton, California and San
Antonio, Texas.1
The results of the pretest led to some further
changes in the scale. Chief among these changes was the
1Colton is a small community near San Bernardino, and
was chosen for its proximity to San Bernardino (where the
actual data were to be collected), and its small size
(which made drawing a sample easier). San Antonio was
chosen because in some ways its population resembled that
of Austin, the other community where data were to be
collected.
75
decision to include nineteen items in the final
instrument. This came about because the three items
devised to measure the Social Play dimension consistently
performed unsatisfactorily in factor analyses and
reliability tests. A fourth item, which experience had
taught us was quite reliable but also quite similar
semantically to another item already in the scale, was
added to assist the measurement of this dimension. The
nineteen items included in the final instrument, and the
dimensions of dependency they were designed to measure,
appear below:
Social Understanding:
Stay on top of what is happening in the
community
Find out how the country is doing
Keep up with world events
Self Understanding:
Gain insight into why you do some of the things
that you do
Imagine what you'll be like as you grow older
Observe how others cope with problems or
situations like yours
Interaction Orientation:
Discover better ways to communicate
Think about how to act with friends, relatives,
or people you work with
Get ideas about how to approach others in
important or difficult situations
Action Orientation:
Decide where to go for services such as health,
financial, or household
Figure out what to buy
Plan where to go for evening and weekend
activities
Social Play:
Provide amusing things to share with others,
such as jokes and cartoons
Give you something to do with your friends
Have fun with family and friends
76
Be a part of events that you enjoy without
having to be there
Solitary Play:
Unwind after a hard day or week
Relax when you are by yourself
Have something to do when nobody else is around
A Scale to Measure Perceived Threat
Concurrent to the development of the MSD scale
described above, other parts of the final instrument were
being discussed. My interest in the concept of threat
perceptions had not developed into the conceptualization
found in Chapter Two by the time this research got
underway. I was nonetheless interested in taking this
opportunity to measure the concept. I shared this
interest with Prof. Ball-Rokeach, who directed me to some
of the literature previously reviewed. In the fall and
winter of 1988-89 we became dissatisfied with the
measurement of this concept (or the lack of such) in
previous research, and Prof. Ball-Rokeach drafted a set of
I
items which we began to test along with the dependency
items. These eight items proved quite reliable in early
pretests, and received little revision before the final
instrument was constructed. Our intention was to measure
the perception of threat at two levels, the personal and
the social. Additionally, distinct types of threats, from
economic problems to natural disasters, were to be
included. The following are the eight items, which were
77
followed by a Likert-style response format from 1
(Extremely Concerned) to 4 (Not at all Concerned):
Think about things that could threaten your
well-being. How concerned are you that the
events listed below could happen to YOU or to
members of YOUR family? Please circle the
appropriate number for each item.
I AM (Extremely Concerned . . . Not at all
Concerned)
ABOUT:
being mugged or robbed
a natural disaster (e.g., earthquake or flood)
that causes damage in my area
being unable to afford the rising cost of
medical care
a nuclear disaster or accident
an economic crisis (e.g., inflation or a serious
recession)
getting a serious illness
finding out that I live in an area that has
unsafe water, land, or air
losing my job
(See Appendix A.)
Measures of Media Use
There have been a variety of ways proposed to measure
the amount of time people spend using various media.
Comparability between studies is often made complicated by
the subtle and sometimes severe differences in the ways
this problem is approached. Two considerations dominated
our decisions in this matter, (1) constructing items that
were similar enough to one another to allow comparison
within the study, and (2) using units of measure which
most resembled the time periods devoted to each medium.
The first concern produced four measures which were
78
similar in wording and in response format. The second led
us to use an average week rather than an average day as
the basis of measurement for magazine reading, and a daily
average for the other three media. Also, in recognition
of different patterns of use for each medium, the print
media were measured explicitly in terms of hours and
minutes, while the use measures of the broadcast media
prompted respondents only by hours. This was done as a
result of pretests in which respondents seemed hard
pressed to specify the amount of time devoted to broadcast
media with precision, while most did not use print media
in blocks of time which could be measured in hours.
The following four questions captured the frequency
of use of the mass media:
Approximately how many minutes and hours do you
spend reading magazines each week?
_____ Hours _____ Minutes
Approximately how many hours of television do
you watch on an average day?
Write your answer here: _____
Approximately how many minutes and hours do you
spend reading a newspaper each day?
_____ Hours _____ Minutes
Approximately how many hours of radio do you
listen to on an average day?
Write your answer here: _____
Other Items on the Survey Instrument
Aside from those scales and questions discussed
above, there were several other distinct parts of the
79
instrument. Some "yes or no" questions asked about use of
cable television, public access television, VCR's, public
television, and public radio. One question addressed
perceptions of ambiguity and another asked how respondents
dealt with "the confusing aspects" of modern life. Each
of the amount of use questions described above was
followed by a question asking which magazines, TV
stations, newspapers, and radio stations respondents
"really looked forward" to using. Another series of
questions called on respondents to estimate the number of
media outlets of various kinds available in their
community (e.g. the number of television stations or the
number of video rental stores).
Another part of the instrument featured the Rokeach
Values Scale (Rokeach, 1982; c.f. Rokeach, 1973), and
questions designed to test ideas regarding value framing
(Ball-Rokeach and Rokeach, 1987; Guthrie, 1988; Ball-
Rokeach et al.. 1990). Finally, we asked questions
regarding the languages spoken at home, number of people
in the household, age, employment status and occupation,
years of residence in the community, sex, ethnicity,
support for interest groups or community groups, income,
and education. (The complete instrument appears in
Appendix A).
80
The Instrument in Two Communities
Part of the complexity of the present study involved
its simultaneous execution in two distant communities. In
California the city of San Bernardino was chosen as the
population from which a sample would be drawn. This
Southern California city, with a population in 1990 of
approximately 130,000, was chosen for three reasons, (1)
it was close enough to Los Angeles to be easily studied on
the variety of levels pertinent to this study, (2) it was
far enough from Los Angeles to enjoy its own distinct
media system, i.e. a large-circulation newspaper (The
Sian), network television affiliates (mountains separate
San Bernardino from Los Angeles, making broadcast signals
difficult to receive from one city to the next), local
magazines, and local radio stations, and (3) it was in a
telephone area code that was considered "local" from
university telephones, making telephone survey research
easier and less costly. In short, San Bernardino provided
the advantages of easy access and a community that was
distinct from the highly unusual media system that Los
Angeles possesses.
In Texas the city of Austin was chosen. Professor
August Grant of the University of Texas, Austin,
coordinated his data collection with ours in California,
but he was not faced with a local community with a highly
unusual media system. While Austin has its own character
as a state capital and the site of a large university, it
has a diverse local economy and its size was comparable to
San Bernardino. We felt that the two cities would offer
reasonable grounds for comparison and contrast when the
data were collected.
By the same token, the research interests of the
groups in each state were not entirely similar. Above I
have described elements which were common to the
instruments distributed in both cities, but there were
differences in the instruments as well. In Austin, the
media for which respondents were prompted with the
nineteen dependency items also included public television,
cable television, and VCR's. These three media were
grouped together as the four others were, and the
respondents were asked to run through the dependency scale
a second time, now prompted by these media. Also, several
other use measures were utilized in Austin. Besides those
detailed above, respondents were asked to report the
number of hours and minutes they spent the previous day
using the media in question. The "previous day" measure
has been used in prior research and is considered by some
a more valid measure of day-to-day use in a large sample
(since the surveys were mailed over the course of a week,
not all respondents were replying on the same day of the
82
week, thus further randomizing the responses to this
question).
The Rokeach Values Survey and the questions regarding
value framing were not included in the Austin survey
instrument. As a result, the length of time needed to
complete the Austin instrument was probably less than that
required by the San Bernardino instrument. This time
difference is somewhat moderated by the extra pass through
the MSD scale required of Austin respondents.
One other difference in the instruments between the
two communities must be noted. In Austin, the threat
scale offered responses ranging from 1 to 5, whereas in
San Bernardino the range was 1 to 4. These differences
necessitated a conversion of responses to normalized "z-
scores" before the data sets could be compared and, then,
combined for certain analyses. This conversion should be
kept in mind when the mean scores are examined (see
Appendix B for a copy of the instrument used in Austin).
To summarize, two survey instruments were constructed
with broadly similar aims and items. Each contained a
nineteen-item scale designed to measure the intensity of
individual media system dependency relations which was
pretested for over six months. In each instrument
intensity was measured for newspapers, radio, magazines,
and television. Each instrument also contained a scale to
83
measure the perception of threat in the respondent's
environment. Each also contained identical questions
regarding the amount of time devoted to use of four mass
media. Each also contained identical questions regarding
the respondent's background and other demographic data.
While the instruments differed in some important ways
(which will receive more attention below), the essential
purposes of the research at the individual level were met
by the instruments.
The Sample and Design
It may be inferred from the description of the
instruments above that the respondents were given a
somewhat lengthy and involved document. Anticipating some
resistance to the completion of such a formidable
document, we decided to design a method of identifying
participants for the pencil-and-paper instrument which
would involve a prior personal commitment to take part in
the research. We decided to call random telephone numbers
in each city, and then randomly select the person in each
household whom we would ask to complete the survey. Two
aims were accomplished through this method, (1) as
mentioned, we felt that if we specifically asked each
respondent to complete and return the instrument when it
arrived in the mail that the response rate would be higher
than if we mailed the instrument "cold," and (2) we could
84
target specific members of each household on the telephone
(e.g. the oldest male, etc.) and thus make certain that
the instrument was addressed to a sample that was
relatively representative of the population.
On the day following the telephone calls, surveys
were sent to those who had agreed to take part the night
before. Of course, once the surveys left our offices it
was up to the respondents to follow through on their
commitment. As will be seen below, it is difficult to
assess the success of this method, particularly in San
Bernardino. We did not have a sample generated without a
telephone call against which we could compare the response
rate of those from whom we obtained the prior commitment.
Our aim was to obtain agreements from about 1000 people
aged sixteen and older in each city, from whom we expected
to receive between 700 and 800 completed surveys (i.e.
response rates of 70% to 80%) in each city.
In order to accomplish this, 5000 telephone numbers
were generated randomly from the telephone exchanges in
each city. This was accomplished using a method described
by Dillman (1978). (1) The number of column inches of
listings in the White Pages was estimated by measuring the
length of a full column and multiplying by the number of
columns per page and the number of pages. (2) Large
sections devoted to non-residences (such as universities,
85
government bureaus, and military bases) were measured and
subtracted from the total. (3) The remaining number of
column inches was divided by 500. The resulting quotient
became the sampling interval. (4) Starting with the first
column on page one, the sampling interval was measured off
and the listing which appeared at the end of the interval
was examined. If the listing was obviously not a
residence, the next residential listing below was
selected. Only the first four digits of the number were
recorded (i.e. the exchange—or prefix—and the fourth
number). (5) Step 4 was repeated until the entire book
was examined (again skipping large sections of non-
residential numbers defined in step 2). The result was
500 randomly selected residential listings. These were
sorted. The frequency of each occurrence of an exchange
and first following digit was noted, and ten times that
number of random "suffixes"—i.e. the last three digits of
the seven digit phone number—were generated by computer.
The result was a list of 5000 telephone numbers
randomly generated with an intention to avoid exchanges
which were not likely to include residences. Of the
sixteen telephone exchanges serving San Bernardino in
1989, four were excluded from the sample entirely. Others
had only one or two of the ten possible nfourth digits"
represented in the sample. While we could not expect to
86
eliminate non-residences from the sample, we could
diminish the number we could expect to encounter while
taking advantage of random digit dialing (i.e. reaching
unlisted numbers and new numbers not yet in the directory)
(Dillman, 1979). This procedure was followed in Austin
and in San Bernardino.
On April 2, 1989, we began making the telephone calls
in both cities, and the following morning the first
surveys were sent. The "recruitment" of respondents
continued in earnest for ten days, each day followed by a
wave of surveys sent to those who had agreed. Often
numbers had to be called several times, in evenings and
mornings, before we reached a potential respondent or
learned that the number did not belong to a residence.
These follow-up initial calls lasted until late April. If
a respondent did not return his or her survey instrument
within a month of the initial mailing, a follow-up call
was made to request that the task be completed. Two weeks
later, post cards were sent to those who still had not
replied reminding them of our request. Finally, a third
call was made two weeks later to those who still had not
responded. After the third follow-up—which represented
our fifth contact (counting the initial call and the
survey itself)—attempts to obtain a completed instrument
87
Table 4.1 Response rates in Austin and San Bernardino.
Telephone Identification of Respondents
Randomly
Generated Ineligible
Site Numbers Numbers2 Refusals
Agreements
Austin 5000 3603 475 922
San Bernardino 5000 3359 742 899
Responses of Those Agreeing in the Telephone Stage
Surveys Valid surveys Response
Site Sent returned rate
Austin 922 653 70.8%
San Bernardino 899 469 52.2%
2Including businesses, numbers not in service, etc.
88
were generally abandoned. This process continued until
late in May.
Table 4.1 presents a breakdown of the response rates
in San Bernardino and Austin. Plainly, the rate in
Austin—70.8%—was much better than the 52.2% response rate
in San Bernardino. There are several potential reasons
for this. (1) Residents of San Bernardino may have been
more intimidated by the length and size of the survey
instrument than those in Austin since inclusion of the
Rokeach Values Scale had mandated a larger format for the
California document. On the other hand, some few
respondents told us after completing the survey that they
enjoyed the values survey the most and were glad of the
opportunity to complete it.3 (2) Residents of Southern
California may be more wary of telephone solicitations and
uninvited mail than those of Austin. (3) Rivalry between
the cities of San Bernardino and Los Angeles may have led
some to discard the survey when they realized its origin.
(4) Rivalry between the University of Southern California
and other area universities—notably the University of
California at Los Angeles (UCLA), the University of
California at Riverside (Riverside is a neighbor city to
San Bernardino) and the California State University at San
3This is not an uncommon reaction to the Rokeach Value
Survey, c.f. Rokeach, 1973.
89
Bernardino—may have further discouraged participation
(some of the people reached by telephone agreed or refused
to participate explicitly referring to USC as the
determining factor).
Whatever the reason, we were left with a somewhat
smaller sample in each city than we desired. In San
Bernardino 469 completed surveys were received. In Austin
653 completed surveys were received. The total sample
when the data sets are combined is 1122.
Tables 4.2, 4.3, and 4.4 present pertinent
information regarding the respondents from both cities.
Some key differences between the two cities' respondents
should be noted. Particularly striking is the
disproportionate number of women who responded to the
survey in San Bernardino (see Table 4.2). The people to
whom the survey was sent were relatively evenly split by
sex, so this result is due to the response rate. The
disparity between the sexes in California is even more
striking when compared to the perfect proportion assembled
in Texas since much of the survey instrument in Austin was
similar to that in San Bernardino. No doubt this
essential difference plays a role in some of the other
differences between the samples to be noted.
The Austin respondents are much better educated than
those in San Bernardino, as demonstrated by the percentage
90
in Austin—over 30%, compared to 14.5% in San
Bernardino—who report more than 16 years of formal
education (see Table 4.2). This may be explained by
Austin's status as a seat of government and the
professional population that such status implies. While
both cities have major universities in and around their
city limits, the state university in San Bernardino does
not offer doctorate degrees (although nearby UC Riverside
does), while the University of Texas in Austin has a wide
variety of respected graduate programs offering the Ph.D.
Thus not only students, but the faculty at Texas may lead
to a better educated population from which the sample was
drawn. The median income range in Austin is higher than
that in San Bernardino, reflecting this education
difference (see Table 4.4). This income differential is
probably related to the uneven sex distribution in San
Bernardino as well.
Those in San Bernardino tend to have lived there
longer than those in Austin (see Table 4.3). This too may
be related to the government and university populations in
Austin. San Bernardino respondents were more likely to be
unemployed and not looking for work than those in Austin,
but this may be due to a larger proportion of housewives
responding in San Bernardino (see Table 4.3). The ethnic
distribution is wider in San Bernardino than in Austin,
9i
Table 4.2 Summary statistics for age, sex, and education,
for respondents from Austin and San Bernardino.
AUSTIN SAN BERNARDINO
Frequency
16-27 179
28-35 146
36-50 168
51-90 126
No resoonse 34
Age
Percent
27.4
22.4
25.7
19.3
5.2
Frequency
107
94
102
136
30
Percent
22.8
20.0
21.7
29.0
6.4
TOTALS 653 100.0 469 100.0
Mean: 38.5
F = 13.018,
s.d.: 15.
p < .000
6 Mean: 42.1 s.d.: 17.4
Frequency
Women 326
Men 326
No resoonse 1
Sex
Percent
49.9
49.9
.2
Frequency
278
183
8
Percent
59. 3
39.0
1.7
TOTALS 653 100.0 469 100.0
Chi-square == 11.55, p = .0007
Years Frequency
1-12 129
13-14 131
15-16 190
17-23 201
No resoonse 2
Education
Percent
19.8
20.1
29.1
30.8
.3
Frequency
165
137
89
68
10
Percent
35.2
29.2
19. 0
14.5
2.1
TOTALS 653 100. 0 469 100.0
Mean: 15.3
Chi-square =
s.d.: 3
= 129.06, p < .000
Mean: 13.9 s.d.: 3.2
92
Table 4.3 Summary statistics for years of residence in the
community, ethnicity, and employment status, for
respondents from Austin and San Bernardino.
AUSTIN SAN BERNARDINO
Years of Residence in the Community
0-4
5-10
11-24
25-74
No response
Frequency
205
174
163
104
7
Percent
31.4
26.6
25.0
15.9
1.1
Frequency Percent
109
69
110
162
19
23.9
14.1
23.5
34.5
4.1
TOTALS 653 100.0 469 100.0
Median: 8.5 Mode: 5 Median: 18 Mode:
F = 45.98, p < .000
Ethnicity
Frequency Percent Frequency Percen'
Asian 13 2.0 15 3.2
Black 32 4.9 38 8.1
Caucasian 492 75.3 293 62.5
Hispanic 61 9.3 69 14.7
Nat. Amer. 39 6.0 36 7.7
Other 11 1.7 7 1.5
No resoonse 5 .8 11 2.3
TOTALS 653 100.0 469 100.0
Chi-square = 20.57, p = .001
Employment Status
Frequency Percent Frequency Percen'
Unemployed,
not looking
for work 127 19.4 130 27.7
Unemployed 35 5.4 32 6.8
Part-time 94 14.4 66 14.1
Full-time 386 59.1 227 48.4
No resoonse 11 1.7 14 3.0
TOTALS 653 100.0 469 100. 0
Chi-square = 14.87, p < .002
Table 4.4 Summary statistics for income for respondents
from Austin and San Bernardino •
AUSTIN SAN BERNARDINO
Income
Frequency Percent Frequency Percent
Under $10,000 70 10.7 70 14.9
$10,000-19,999 105 16.1 87 18.6
$20,000-29,999 100 15.3 86 18.3
$30,000-39,999 87 13.3 83 17.7
$40,000-49,999 75 11.5 39 8.3
$50,000-59,999 64 9.8 27 5.8
$60,000-69,999 43 6.6 24 5.2
$70,000-79,999 27 4.1 11 2.3
Over $80,000 56 8.6 17 3.6
Missina 26 4.0 25 5.3
TOTALS 653 100.0 469 100.0
Median: $20,000-29,999 Median: $10,000-19,999
Chi-square = 31.87, p = .0001
94
Table 4.5 Summary
years of
combined
statistics for
residence in
data set.
age, sex, education, and
the community for the
16-27
28-35
36-50
51-90
No response
Age
Frequency
286
240
270
262
64
Percent
25.5
21.4
24.1
23.4
5.7
TOTALS 1122 100.0
Mean: 40 s.d.: 16.4
Women
Men
No response
Sex
Frequency
604
509
9
Percent
53.8
45.4
.8
TOTALS 1122 100.0
Years
1-12
13-14
15-16
17-23
No response
Education
Frequency Percent
294 26.2
268 23.9
279 24.9
269 24.0
12 1.1
TOTALS 1122 100.0
Mean: 14.7 s.d.: 3.14
0-4
5-10
11-24
25-74
No response
Years of Residence in the Community
Frequency Percent
317 28.3
240 21.4
273 24.3
266 23.7
26 2.3
TOTALS 1122 100. 0
Median: 11 Mode: 5
95
Table 4.6 Summary statistics for ethnicity, employment
status, and income for the combined data set.
Asian
Black
Caucasian
Hispanic
Native American
Other
No response_____
TOTALS
Unemployed,
not looking
for work
Unemployed
Part-time
Full-time
No response
TOTALS
Under $10,000
$10,000-19,999
$20,000-29,999
$30,000-39,999
$40,000-49,999
$50,000-59,999
$60,000-69,999
$70,000-79,999
Over $80,000
Missing________
Ethnicity
Frequency
28
70
785
130
75
18
16
Percent
2.5
6.2
70.0
11.6
6.7
1.6
1.4
1122 100.0
Employment Status
Frequency Percent
257
67
160
613
25
22.9
6.0
14.3
54 . 6
2.2
1122
Income
Frequency
140
192
186
170
114
91
67
38
73
51________
100.0
Percent
12.5
17.0
16
15
10 ,
8 ,
6
3,
6,
4,
6
2
2
1
0
4
5
5
TOTALS 1122 100. 0
Median: $20,000-29,999
96
but both cities are represented mostly by whites, secondly
by Hispanics (see Table 4.3). The significant differences
between the samples suggests that all subsequent analyses
should be performed, where possible, on data sets
consisting of each city's sample as well as the combined
data set.
Tables 4.5 and 4.6 offer a summary of the sample when
the data sets are combined.
Analyzing the Data
The analysis of the validity and reliability of the
threat and MSD scales will proceed along two paths. To
determine the internal reliability of the scales,
Cronbach's alpha will be calculated. Cronbach's alpha is
a test of the intercorrelation of the items. If, in fact,
the items in a scale are measuring different facets of a
single concept, for the most part they ought to correlate
quite highly with one another. For the threat scale, and
for the three-item dependency scales for all six
substantive motivations and for all four media tested in
both cities, Cronbach's alpha will be computed for the
combined data set and for each city individually.
While the reliability test is important, it is also
helpful to test the conceptual validity of the scales by
factor analysis. For the threat scale and the MSD scale,
exploratory factor analyses will be performed to test the
97
dimensionality of the scales. The threat scale is
intended to be unidimensional, and orthogonal factor
analysis will be performed to test this assumption. The
18-item dependency scale is intended as a measure of six
underlying, distinct factors for each medium. The
performance of these 18 items in this regard will be
tested for each medium. Since the MSD scale is expected
to yield a definite number of dimensions—dimensions
defined by certain specified items in the scale—it is
possible to perform a confirmatory factor analysis as
well. In short, the exploratory factor analysis will test
the straightforward assumption that the MSD scale is
multidimensional, while confirmatory factor analysis will
test the more rigorous theoretical claim of specific
factors.
The first research question, regarding the bivariate
relationships between threat perceptions and certain
demographic variables, will be addressed using correlation
analyses and analyses of variance. Age, education, and
years of residence in one's community are all at least
interval level data, and may be correlated with threat
using Pearson's r. Income, sex, ethnicity, and employment
status are all measured as ordinal or categorical
variables in this survey, and will be related to threat
perceptions through analysis of variance. Differences in
98
threat perceptions between residents of San Bernardino and
Austin may not be measured with an analysis of variance
since their scales were of different lengths. The means
and distributions of the responses from each city will be
compared, however, and any apparent differences noted.
The second research question, regarding the empirical
dimensions in the data regarding media system dependency
relations, will have been addressed in the scale analyses
described above.
The first hypothesis will be tested in two ways, due
to the different levels of data for measuring dependency
relations that are available in these data. Hla and Hlb
involve measures of the intensity of dependency relations,
which are interval level variables in these data. In
order to test the hypothesis that these variables are
related to threat perceptions in a non-linear
form—specifically a parabolic or bell-shaped curve—a
regression analysis will be performed. Since I am
hypothesizing a parabolic form, the equation to which the
data should conform is:
y = B0 + B^2
While the parabolic equation defines a type of nonlinear
relationship, if the square of x is substituted for x in
the equation, one is left with the linear equation:
y = B0 + B,X'
99
where X' is the squared values of all observed values for
X. The parameters in this equation may be estimated using
linear regression. A transformation of the values of
threat perceptions—the independent variable in the
hypotheses dealing with parabolic equations, i.e. Hla and
Hlb—will be performed, and regression analysis will test
the suitability of the parabolic equation to the data.
Hypothesis 1 specifies a parabola with the shape of an
inverted U, or a bell. In order to describe this
particular curve, the Beta term associated with the
squared term in the parabolic equation must be negative.
Therefore, the following equation is the only form which
will support my hypothesis:
y = B0 + (~‘ B1) X'
Regression may not be performed on categorical level
dependent variable. The measures of substantive scope and
media scope will be secondary measures derived from
responses to the MSD scale, placing respondents in
categories of high, medium, and low breadths of scope in
these variables. High substantive scope will be
operationalized as intensity of dependency above one
standard deviation in two or more substantive motivations
in a given medium. Low substantive scope will be
operationalized as intensity of dependency higher than the
mean in no more than one substantive motivation in a given
100
medium. Moderate substantive scope is all other
combination of intensity scores. For instance, if a
respondent reported intensity beyond one standard
deviation in both the social understanding and social play
substantive motivations for radio, he or she will be
included in the high substantive scope group for that
medium. High media scope will be operationalized as
intensity at least one standard deviation higher than the
mean in a given substantive motivation in three or four
media. Low media scope will be operationalized as
intensity lower than the mean in all media in a given
substantive motivation. If one reported intensity beyond
one standard deviation in social understanding for
television, newspapers, and radio, he or she will be in
the high media scope group for social understanding.
The categorical measures of the dependency scope
variables suggest one-way analysis of variance and the
Scheffe' test as the test of hypotheses Hlc and Hid. Two
observations will have to be made to support these
hypotheses: (1) The average perceptions of threat in the
three groups must rise and fall as hypothesized, and (2)
the Scheffe* test must demonstrate that the differences
between the averages for each group are statistically
significant (p < .05).
101
Hypotheses 2 and 3 may be tested using multiple
regression, with certain categorical variables controlled
through dummy codes. Likewise, the third and fourth
research questions may be addressed through this method.
While the dummy codes and, in some cases, the treatment of
variables that are ordinal in strict terms as if they are
interval (i.e. income) will violate certain assumptions of
the linear model, it is hoped that the method is robust
enough to withstand this challenge in exchange for
parsimonious tests of these hypotheses. The alternative
is a battery of tests using diverse combinations of
control variables. Multiple regression equations will
also allow for examination of the results of the tests of
Hla and Hlb when other variables are controlled; if the
earlier hypotheses are supported, I can examine whether
the non-linear equation is still the best fit when other
variables are controlled. If no support for the non
linear equation is found in the simple regression
analysis, perhaps some support will emerge after
controlling other variables.
The following chapter will present the results of
these analyses, and a subsequent chapter will discuss the
findings.
Chapter Five: Results
Threat Perceptions
In order to investigate the bivariate relations
between perceptions of threat and demographic variables it
is necessary to test the reliability of the threat scale
and determine whether these items constitute a reliable,
unidimensional scale. Two statistics are available for
this purpose, Cronbach’s alpha and the eigenvalues in
factor analysis. In the combined data set, Cronbach's
alpha for the eight item threat scale is .8192. In
Austin, Cronbach's alpha is .8225, while in San Bernardino
it is .8142. Not only would the scale appear to be
reliable, but consistent in its performance in this
regard. In no analysis could the alpha be improved by
omitting any of the items.
To further test the dimensionality of the scale, a
factor analysis was performed. The presumption of a
unidimensional scale is that no more than one factor will
achieve an eigenvalue of 1, and that a large amount of
variance will be explained by the first factor. The
results of an orthogonal factor analysis of the threat
items in the combined data set show that the first factor
has an eigenvalue of 3.5, and explains 44.3% of the
variance. No other factor has an eigenvalue greater than
one, and the one-factor solution prevents the calculation
■Table 5.1 Reliability analysis of Threat Perception j
I Scale. |
! ;
RELIABILITY COEFFICIENTS
Combined sample Austin sample San Bernardino sample .
N - 1037 N = 619 N = 418 ]
ALPHA = .8192 ALPHA = .8225 ALPHA = .8142 I
FACTOR ANALYSIS
Listwise deletion of cases with missing values
Principal-Components Analysis
Final Statistics:
Variable Communality
Being mugged or robbed .43965
A natural disaster that causes damage .45605
'Being unable to afford medical care .38478
■A nuclear disaster or accident .51155
An economic crisis .44970
iGetting a serious illness .47245
'Finding out that an area has unsafe water, etc. .48976
jLosing my job .33793
;Eigenstructure:
I
'Factor Eigenvalue Pet of Var
1 3.54187 44.3
104
of factor score coefficients. The complete results of the
reliability analyses of the threat scale appear in Table
5.1.
I conclude that the threat scale performs reliably in
both cities as a unidimensional measure of the perception
of threat in the environment. By summing the scores of
each respondent on all eight items, then dividing that sum
by eight, a single "threat score" may be calculated.1 It
is this score which will be compared with other variables
to examine the relationships between threat and
demographic characteristics.
First the comparison of each city's threat scores
will be made, however. In Austin, the mean threat score
was 2.9l2 and the standard deviation was .813. This
distribution is near normal, with a kurtosis of -.08 and
skewness of -.2l.3 In San Bernardino, the mean threat
score was 2.56 and the standard deviation was 1.18. This
1It will be recalled from the previous chapter that
the summation involved is a summation of z-scores. due to
the incompatibility of the scales from the two sites.
2These threat scores are uncorrected, i.e. lower
values indicate higher perceptions of threat.
Measures of kurtosis (the tendency of the scores to
cluster about the mean) and skewness (the tendency of the
scores to be higher or lower than the midpoint of the
scale) are such that a perfectly normal distribution would
have scores of 0. As will be seen, these statistics
provide a means of comparing the two cities' distributions
for the threat perception scale.
105
distribution is clustered remarkably about the mean, with
a kurtosis of 13.46, and is positively skewed far more
than Austin's is negatively skewed; skewness in San
Bernardino was 3.00. In San Bernardino, the mean score
shows higher threat perceptions than in Austin, and a much
greater proportion of scores clustered about the mean than
in Austin. It would appear that the actual mean
perception in the San Bernardino sample is slightly higher
than in Austin, but that more of the California
respondents are likely to report that mean score than
those in Texas.
The first research question simply seeks to examine
the bivariate relationships between seven demographic
variables and perceptions of threat. Three of these
variables, (1) age, (2) education, and (3) years of
residence in one's community, are made up of interval
level data and may be related to threat through
correlation analysis. The other four are reported by
categories, (1) sex, (2) income,4 (3) employment status,
and (4) ethnicity. These variables may be examined
through analyses of variance, comparing the means of
threat perceptions within each category. In these
bivariate analyses, no controls are introduced for the
4Income was reported in ranges of $10,000 (See
Appendix A).
106
Table 5.2 Bivariate correlations between threat
perceptions and age, education, and years of
residence in the community.
Threat and Age
Combined Sample r = -.06
Austin------------ r = -.03
San Bernardino r = -.10*
Threat and Education
Combined Sample --- r = -.17***
Austin--------------r = -.18***
San Bernardino r = -.16***
Threat and Years of Residence in the Community
Combined Sample --- r = .00
Austin--------------r = .03
San Bernardino ---- r = -.03
* = p < .05
** = p < .01
*** = p < .001
107
other variables. When tests of the hypotheses in Chapter
3 are performed, more care will be taken to account for
the interaction between variables. For now, I will
present only the most elementary associations in the
combined sample as well as the separate city samples.
The correlation analyses (see Table 5.2) show that
when no controls are introduced threat is significantly
related only to education. The less educated one is, the
more he or she perceives the environment to be
threatening.5 The magnitude of the correlation between
education and threat is small, accounting for only 3% of
the variance between the two. Although the correlation
between threat and age is moderately significant (p =
-.043) in San Bernardino, the sample size makes this level
of significance relatively easy to accomplish. The
analyses of variance demonstrate significant differences
between several subgroups of each site’s population.
Particularly striking is the combined sample's employment
status result. The unemployed report levels of threat
perception that are significantly higher than those
reported by any other group (see Table 5.5). The
difference between sexes in the combined sample is highly
5The original responses to the threat scale in these
and in all subsequent analyses were reversed so that
higher scores indicate higher perceptions of threat. (See
Appendix A for the threat scale as it appeared in the
survey instrument.)
108
Table 5
Women
Men
Women
Men
Women
Men
3 Bivariate analysis of threat perceptions and
sex. Analysis of variance.
Threat perceptions by Sex
Combined sample
N Mean F = 36.043 p < .000
540 .11 d.f. = 1
491 - .13
Threat perceptions by Sex
Austin
N Mean F = 21.452 p < .000
300 .12 d.f. = 1
318 - .12
Threat perceptions by Sex
San Bernardino
N Mean F = 14.975 p < .000
240 .10 d.f. = 1
173 - .15
109
N Mean
1. Under $10,000 129 .2174
2. $10,000-19,999 172 .0738
3 . $20,000-29,999 168 .0143
4. $30,000-39,999 162 . 0117
5. $40,000-49,999 106 -.0036
6. $50,000-59,999 90 -.1342
7. $60,000-69,999 64 -.2635
8. $70,000-79,999 36 -.2605
9. Over $80,000 71 -.2145
Table 5.4 Bivariate analysis of threat perceptions by
income. Scheffe test.
Threat perceptions by Income
Combined sample
Income levels
1 2 3 4 5 6 7 8 9
* *
* =
p < .05
Threat perceptions by Income
Austin
Income levels
N Mean
1. Under $10,000 68 .1409
2. $10,000-19,999 99 .0643
3. $20,000-29,999 92 . 0919
4. $30,000-39,999 83 .0741
5. $40,000-49,999 70 -.0074
6. $50,000-59,999 64 -.1030
7. $60,000-69,999 42 -.2382
8. $70,000-79,999 25 -.2495
9. Over $80,000 55 -.2413
(No significant differences)
Table 5.4 (cont.) Bivariate analysis of threat
perceptions by income. Scheffe test.
Threat perceptions by Income
San Bernardino
Income levels
N Mean
1. Under $10,000 61 .3027
2. $10,000-19,999 73 .0867
3. $20,000-29,999 76 -.0796
4. $30,000-39,999 79 -.0539
5. $40,000-49,999 36 .0039
6. $50,000-59,999 26 -.2110
7. $60,000-69,999 22 -.3119
8. $70,000-79,999 11 -.2853
9. Over $80,000 16 -.1224
(No significant differences)
Table 5.5 Bivariate analysis of threat perceptions by
employment status. Scheffe test.
Threat perceptions by Employment status
Combined sample
111
Status
1. Unemployed— not looking
2. Unemployed— looking for work
3. Employed part time
4. Employed full time
N Mean 1 2 3 4
207 -.0946
64 .3654
* * *
155 .0582
594 -.0241
* =
= p < .05
Threat perceptions by Employment status
Austin
Status
1. Unemployed— not looking
2. Unemployed— looking for work
3. Employed part time
4. Employed full time
N Mean 1
113 -.1404
33 . 3126
*
92 . 0007
373 . 0171
* =
p < . 05
Threat perceptions by Employment status
San Bernardino
Status
1. Unemployed— not looking
2. Unemployed— looking for work
3. Employed part time
4. Employed full time
N Mean 1 2 3 4
94 -.0395
31 .4215
* *
63 .1421
221 -.0936
* =
p < . 05
112
Table 5.6 Bivariate analysis of threat perceptions by
ethnicity. Scheffe test.
Threat perceptions by Ethnicity
Combined sample
Ethnicity
N Mean 1 2 3 4 5 6
1. Asian 26 .2879
2. Black 58 .1374
3. Caucasian 739 —.1082
* *
4. Hispanic 121 .3495
5. Native American 62 .2093
6. Other 18 .2742
* = p < .05
Threat perceptions by Ethnicity
Austin
Ethnicity
N Mean 1 2 3 4 5 6
1. Asian 13 .3145
2. Black 27 .4080
3. Caucasian 468 -.1084
* *
4. Hispanic 61 .4309
5. Native American 34 .1937
6. other 11 .0133
p < .05
Threat perceptions by Ethnicity
San Bernardino
N Mean
1. Asian 13 .2614
2. Black 31 -.0983
3. Caucasian 271 -.1078
4. Hispanic 60 .2668
5. Native American 28 .2282
6. Other 7 .6841
Ethnicity
1 2 3 4 5 6
* = p < .05
113!
significant as well, with women reporting higher
perceptions of threat than men (see Table 5.3). The
analysis of ethnicity shows that those who identified
themselves in the combined sample as Hispanic or as Native
American have perceptions of threat that are significantly
higher than white respondents (see Table 5.6). The
analysis of income shows that the lowest income category
(those households earning less than $10,000 in 1988)
perceive more threat than two of the three highest income
categories (see Table 5.4). The analyses of variance for
individual cities shows a different pattern. In Austin,
the unemployed perceive more threat than the unemployed
who are not looking for work, but not significantly more
than any other group (see Table 5.5). Also, in Austin it
is both blacks and Hispanics who demonstrate higher threat
perceptions than whites (see Table 5.6). Austin shows a
significant difference between the sexes, with women—as in
the combined sample—reporting higher perceptions of threat
than men (see Table 5.3).
In San Bernardino, the analyses of variance show the
unemployed reporting higher threat perceptions than the
full time employees and those unemployed who are not
looking for work, but no significant difference between
the unemployed and part time employees (see Table 5.5).
Hispanics perceive significantly more threat than do
114
whites, but no other main effect for ethnicity is observed
(see Table 5.6). Finally, as in all other analyses, women
perceive significantly more threat than do men in San
Bernardino (see Table 5.3).
To summarize the bivariate analyses: The less
educated, women, Hispanics, and the unemployed
consistently report higher perceptions of threat than the
better educated, men, and most other categories of
employment status. While the specific patterns of
relationships between these variables change between sites
and in the combined sample, on the whole it is clear that
members of society who are in some way unempowered tend to
perceive their environment as more threatening than
educated, employed, white men. In neither city is the
difference between income categories significant. The
difference noticed in the combined sample could be the
result of error (significance at p < .05 with eight
independent tests is not terrifically difficult to achieve
by chance alone), but it is consistent with the general
pattern of results.
Media System Dependency Relations
The second research question (see Chapter 3) concerns
the dimensions of dependency relations observed in this
data set. The introduction of the scale to measure
individual level intensity of dependency relations (Ball-
115
Rokeach and Grant, 1991) calls for analysis of the scale's
reliability, as well as its success in discriminating
among the six substantive motivations (see Chapter 1). To
explore this research question, three methods were used.
Cronbach's alpha, as with the threat scale above, measured
the reliability of the six three-item scales which
together comprise the 18-item dependency scale. The alpha
statistic, however, does not provide a measure of the
effectiveness of the method of measuring these six
dimensions simultaneously in one 18-item scale. Factor
analysis using oblique rotation (allowing factors to
correlate)6 allows for the most exploratory analysis of
the underlying factor structure in the 18-item scale. A
more rigorous test of the structure of the responses is
provided by a confirmatory factor analysis using a
structural equation model.
Before turning to the minutiae of these analyses,
however, one important decision had to be made. It will
be recalled that the survey instrument in both cities
actually contained 19 items to measure dependency, due to
6Ball-Rokeach et al. (1984) suggest that the
substantive motivations are not mutually exclusive,
although collectively they are held to be exhaustive of
motivations involved in media system dependency relations.
The assertion that these concepts are not exclusive of one
another suggests that their measures ought to correlate
somewhat while still discriminating to some significant
degree.
lie
the inclusion of four social play items (see Chapter 4).
In order to achieve consistency in the number of items per
measure of each substantive motivation, it was necessary
to analyze these data and decide which social play item
would be discarded from the data. The item that was
discarded was "Provide amusing things to share with
others, such as jokes and cartoons." Factor analyses of
the nineteen items showed that this item never once loaded
with another social play item in any medium. While it
will be seen below that social play remains a dimension
that is measured weakly by the three items which comprise
the present scale, that weakness is in no way relieved by
this item. Reliability analyses—the inter-item
correlations—support this decision as well, in that the
item excluded has the lowest correlations across all media
with the other social play items. In all subsequent
analyses of dependency responses, this item is excluded.
Each analysis of the dependency data was conducted
separately for each medium, and will be discussed
separately as well. Some key differences in the scale's
performance across the various media examined will be
noted, and must be kept in mind during subsequent
analyses. The first issue to address is the reliability
of the three-item scales in each substantive motivation.
117
Newspapers
The reliability of the scale to measure the intensity
of dependency relations across the six substantive
motivations for the newspaper medium is uneven (see Table
5.7). For the combined sample, Cronbach's alpha is under
.8 for four of the three-item scales (although the scale
for interaction orientation is very close to .8). The
results in the Austin sample are virtually identical to
those of the combined sample. In San Bernardino, the
reliability is somewhat higher, with one additional scale
crossing the .8 threshold and each other scale performing
marginally better. The social play scale remains an
unstable measure in all samples, although the newspaper
medium must be seen as particularly unsuited to this
motivation and, therefore, perhaps not the best test case
of this scale.
Radio
In the combined sample, the reliability of the scales
for radio is very similar to that for newspapers. The
same two scales show reliability coefficients above .8
(see Table 5.7). The social play scale performs much
better in this instance, perhaps reflecting the
suitability of this medium to the goal of social play. In
Austin, the results are again similar to the combined
sample. San Bernardino does not show any improvement over
118
Table 5.7 Reliability analysis of the individual level
Media System Dependency Scale. Cronbach's
alpha.
Newspapers Sample
Combined Austin San Bernardino
Social understanding .8646 .8629 .8662
Self understanding .7681 .7408 .8016
Interaction orientation .7939 .7948 .7937
Action orientation .7632 .7423 .7865
Social play .6694 .6686 .6712
Solitary play .8205 .8224 .8148
Radio Sample
Combined Austin San Bernardino
Social understanding .8375 .8539 .8125
Self understanding .7670 .7893 .7385
Interaction orientation .7897 .7990 .7749
Action orientation .7774 .7844 .7708
Social play .7884 .7925 .7819
Solitary play .8556 .8543 .8579
Magazines Sample
Combined Austin San Bernardino
Social understanding .7150 .7147 .7137
Self understanding .7722 .7593 .7898
Interaction orientation .8093 .8225 .7891
Action orientation .6805 .6663 .7000
Social play . 6683 .6586 .6859
Solitary play .7990 .8168 .7697
Television Sample
Combined Austin San Bernardino
Social understanding .8164 .8582 .7797
Self understanding .7689 .7688 .7693
Interaction orientation .8099 .8067 .8143
Action orientation .7917 .7666 .8239
Social play .7349 .6923 .7863
Solitary play .8369 .8198 .8571
119
the combined sample, and in fact scales do worse overall
in California than they do in Austin.
Magazines
The scales perform poorly for magazines (see Table
5.7). In the combined sample, only one scale reaches the
.8 alpha level, and many of the others are lower than any
i
seen for newspapers or radio. In Austin, another scale
crosses .8, but the general unsatisfactory trend
continues. The scales for magazines perform worst in San
Bernardino, where no scale has a coefficient higher than
.79.
Television i
In the combined sample, three of the scales I
I
demonstrate reliability beyond the .8 level for television i
j
(see Table 5.7). This is the best performance for the j
i
scales for any medium. In Austin these findings are |
repeated. The San Bernardino data show three coefficients j
I
beyond .8, but only two of the three are the same as those
found in the combined sample and in Austin. Rather than
I crossing that threshold for social understanding, the San !
> I
j ;
l Bernardino data show the action orientation scale j
; performing at this level. !
i i
j Overall, the social understanding and solitary play i
j I
| scales provide the most reliability across all the media.
| Interaction orientation occasionally crosses the .8
120
threshold, and usually remains close to that level.
Social play, as mentioned above, remains disturbingly
unreliable across all media. Attention may now be turned
to the factor analyses with the reliability results in
1 mind.
j
i
| Newspapers
j
; If the scales were perfect, six factors would be
! revealed, each accounting for approximately the same
|
; amount of variance. In fact, in the combined sample three
i
I
j factors with eigenvalues greater than one (and one with an
I
j eigenvalue of .97) are derived from the data. Maximum
1 likelihood extraction saves for rotation the number of
I
I
j factors which contribute to a solution which fits the
|
j correlation matrix in the data at a statistically
j
! significant level. Usually, this means factors with j
| eigenvalues greater than 1 (i.e. factors which explain 10% [
I !
; or more of the variance in the matrix). The three factors ,
I in the combined data for newspapers are clearly
i
interpretable as a social understanding factor, a solitary :
i
play factor, and all other substantive motivations in the j
other factor (see Table 5.87). This same pattern is !
observed in Austin. In San Bernardino, however, there are j
| 7Tables will be presented only for the factor analyses
■ of the combined data set. Differences between the results
j in the combined set and those in each city will be noted
in the text.
121
four significant factors, and a single factor for social
play and action orientation is added to those observed in
the combined data and in Austin.
Radio
Three factors again appear in the combined data set,
with another eigenvalue close to one (.95) (see Table
5.9). These factors are not as "clean" as those for
newspapers, with the exception of a single factor for the
social understanding items. The first factor holds the
interaction orientation and self understanding items and
two action orientation items, while the second holds the j
play items and one action orientation item. In Austin the |
general pattern holds up, although the first factor is j
considerably more muddled, holding all orientation items |
i
as well as two social play items in addition to the self
understanding items. San Bernardino again produces four
i
i
rather than three significant factors, and as with
i
i
newspapers both social understanding and solitary play j
break into distinct factors by themselves. Self *
understanding and interaction orientation items comprise
i
j the third factor, while the first factor is made up of j
! i
I social play and action orientation items.
I
| Magazines
i
i The combined sample produces a three factor solution, ;
but no factor is as clear as the social understanding
122
Table 5.8 Factor analysis of the individual level Media System Dependency scale for
newspapers. Oblique rotation.
Combi ned Data Set
Rotated Pattern Matrix:
Factor 1 Factor 2 Factor 3 CommunaIi ty
How to act with friends .81883 .58311
How to approach others .77902 .57589
Observe others cope .71572 .58262
Have fun with family .66730 .38028
Something with friends .58492 .30822
Imagine old age .58115 .47342
Gain insight into self .56865 .47449
Better communication .55360 .48552
Figure out what to buy .48254 .38711
Where to go for service .35697 .36814
Plan where to go .35352 -.33066 .39822
Enjoy events .34121 .40184
Keep up with world -.90175 .80278
How country is doing -.89108 .78122
On top of community -.61701 .52119
Unwind after hard day -.64130 .50504
Relax by yourself -.97354 .83260
Something to do alone -.59473 .54935
Eigenstructure:
Factor Eigenvalue Pet of Var
1 7.44579 41.4
2 .95687 5.3
3 1.01138 5.6
L
123
Table 5.9 Factor analysis of the individual level Media System Dependency scale for radio.
Oblique rotation.
Combined Data Set
Rotated Pattern Matrix:
How to approach others
How to act with friends
Observe others cope
Gain insight into self
Imagine old age
Better communication
Figure out what to buy
Where to go for service
Relax by yourself
Unwind after hard day
Something to do alone
Something with friends
Have fun with family
Plan where to go
Enjoy events
Keep up with world
How country is doing
On top of community
E i genst ructure:
Factor Eigenvalue Pet of Var
Factor 1
.84453
.80846
.80735
.62161
.59440
.57034
.36902
.34208
Factor 2 Factor 3
.86437
.76278
.75338
.56654
.47755
.40996
-.90557
-.81785
-.61873
Communality
.63705
.58564
.64893
.50144
.44802
.48605
.42950
.41577
.50637
.55830
.61996
.51498
.50637
.50628
.45807
.76111
.70108
.44947
1
2
3
7.97786
1.17118
.74883
44.3
6.5
4.2
124
factors which appear in the newspaper and radio analyses
(see Table 5.10). Each factor is a combination of two
scales; interaction orientation and self understanding in
the first, solitary play and social understanding in the
second, and action orientation and social play in the
third. Austin's data provide a three factor solution as
j well, but this solution is more muddled than the combined
J sample's due to a single factor for solitary play which
throws social understanding, social play, and action !
orientation into the third factor together. San J
Bernardino yields the now-familiar three factor solution, j
but differs from Austin by yielding a distinct social j
understanding factor and combining the self understanding,
interaction orientation, and solitary play scales in the
first factor.
\
| Television
j
| The combined sample's solution for television is
J quite similar to that observed for newspapers in that
1 i
three factors emerge and separate factors for social
| t
i understanding and solitary play are among them (see Table ;
; 5.11). Austin provides the same pattern in its solution.
I San Bernardino again provides three factors, but they are !
different in their composition. The first factor combines '
I
!
self understanding, interaction orientation, and social i
i ;
I play. The second combines solitary play and social
Table 5.10 Factor analysis of the individual level Media System Dependency scale for
magazines. Oblique rotation.
Combined Data Set
Rotated Pattern Matrix:
Factor 1 Factor 2 Factor 3 CommunaIi ty
How to act with friends .83119 .63983
How to approach others .73081 .63514
Observe others cope .72929 .64071
Gain insight into self .63595 .53442
Better communication .61015 .54599
Imagine old age .50998 .47552
Relax by yourself .69547 .58213
Unwind after hard day .65547 .45870
How country is doing .65496 .46487
Keep up with world .63747 .50575
Something to do alone .59426 .52608
Enjoy events .47773 .38950
On top of community .32535 .22744
Plan where to go .64829 .49499
Something with friends .59846 .51575
Have fun with family .36366 .54448 .57487
Figure out what to buy .37519 .43847
Where to go for service .36295 .36686
E i genst rueture:
Factor Eigenvalue Pet of Var
! 1 7.32210 40.7
2 .92325 5.1
3 .77168 4.3
126
understanding. The third factor isolates action
| orientation.
i
J Overall, it is clear that the eighteen items do not
I constitute a unidimensional scale. In no instance does a
J
I
I single factor solution emerge from the maximum likelihood
!
extraction. The scales which proved most reliable—i.e.
i
: social understanding and solitary play—demonstrate the
greatest propensity to fill distinct factors; they do so
in each solution except that of magazines. Self
understanding and interaction orientation consistently
comprise the first factor, sometimes by themselves and
often in combination with one or more other scale items.
The first factor always explains a great deal of variance,
virtually always mOre than 40% of the variance in the
i
correlation matrix. This factor is always comprised of at
least two scales, however, and its heavy explanation of
variance is not by itself an indication of a i
unidimensional scale.
Clearly, however, the eighteen item scale is still
f
beset with problems in certain areas—most notably social j
play and, to a lesser extent, action orientation. It is
left to a confirmatory factor analysis to test the
I
| hypothesis that a solution of six factors, each factor j
!
4 I
; defined by the three items intended to measure the j
| |
l substantive motivations, is a better fit to the data than j
I i
I i
1 _
Table 5.11 Factor analysis of the individual level Media System Dependency scale for
television. Oblique rotation.
Rotated Pattern Matrix:
Combined Data Set
12 7 ,
Factor 1
How to act with friends .85245
How to approach others .83769
Observe others cope .71391
Have fun with family .69163
Better communication .67147
Figure out what to buy .65156
Gain insight into self .64030
Imagine old age .62193
Something with friends .60967
Plan where to go .60248
Where to go for service .55173
How country is doing
Keep up with world
On top of community
Relax by yourself
Something to do alone
Unwind after hard day
Enjoy events
Factor 2 Factor 3
Eigenstructure:
Factor
1
Eigenvalue
8.00895
1.21144
.61513
Pet of Var
44.5
6.7
3.4
-.81432
-.71744
-.48610
-.91067
-.68735
-.65436
-.37301
Communality
.59097
.61533
.58595
.54606
.50098
.47965
.50212
.49785
.44705
.41989
.38661
.78900
.68856
.41551
.76269
.60716
.59871
.40141
J
128
a one-factor solution, and better than a "compromise" J
three factor solution where the social/psychological !
components of the substantive motivations are collapsed
and scales of understanding, orientation, and play are
left, each with six items. The analysis takes the form of
a structural equation model in which the specified factors
are said to be "caused" by specified items in the scales.
This analysis was done only for the combined data set.
In all four media, the best solution by far of the
three described above was the six factor solution (Ball-
Rokeach and Grant, 1991; Horvath, 1990).8 The results
indicated that solutions with between five and seven
dimensions may provide the best fit for some media. This
does not solve all problems, but it does quiet some of the
alarms raised by the earlier factor analyses in which ;
three-factor solutions were most common. In all media,
however, familiar problems appear when the modification
j
indices provided by the LISREL VII analytical software are !
examined. These indices show where improvements in the |
; model may be achieved by changing the causal relationships j
I
, first proposed. In this case, that means allowing items j
1 i
; designed to measure one substantive motivation to
| ®This analysis was performed by August Horvath as part
[ of the research in which the group was involved. His
I conclusions are summarized in an unpublished report
i referenced as Horvath, 1990.
129
contribute to the explanation of one or more other
motivations. Not surprisingly, it is the social play
items which are most commonly recommended for change, but
the first social understanding item ("stay on top of your
community") also appears often, and the solitary play
items perform poorly as well. One final unsurprising
result, given the earlier factor analyses, is that the
interaction orientation items are recommended for
modification and, when freed, load with the self
understanding items.
I conclude that the data provide evidence of six
scales which perform at widely diverse levels of
reliability. My hypotheses are specifically concerned |
I
with the social understanding, self understanding,
i
interaction orientation, and action orientation
motivations. Thus, for present purposes, I am dismayed
i
but not alarmed by the poor performance of the social play
scale. The consistent association of the interaction
orientation and self understanding items in the factor
analyses indicate that subsequent analyses should be
interpreted carefully. The alpha coefficients for these
scales, however, while not outstanding, are near enough to !
130
acceptable standards to justify keeping the scales
separate for the sake of conceptual consistency.9
The scale scores for each substantive motivation will
be constructed by adding the reported scores for each item
! in the scale and dividing by three. In all subsequent
!
analyses, references to variables social understanding.
self understanding, interaction orientation, action j
orientation, social play, and solitary plav refer to the j
i
scale scores for the intensity of dependency relations
j computed in this fashion.
j One assumption of multiple regression is that the
dependent variable has values which are normally
distributed throughout the population. The six dependency
scales described in the preceding paragraph did not always I
i
provide a normal distribution in the data from the two ^
cities. The regression statistic is robust enough to j
i
withstand a considerable amount of violation of the
I
assumption of normality, and certain types of
distributions may be transformed to meet this assumption !
' . i
i if necessary. Unfortunately, the most common problem t
| encountered in the dependency scale was a "spike" at one
or another boundary of the scale. The spikes indicate
9A more complete discussion of the issues involved in
the construction of this scale and the interpretation of
the factor analyses is available in Ball-Rokeach and
Grant, 1991.
131
either that many respondents wished to provide a higher or
lower score on the scale than the scale allowed, or that
an uncommonly large number of respondents were determined
to report the highest or lowest score offered, regardless
of its magnitude.
There is no transformation to solve this sort of
t
i distribution. In order to create variables from these
scales which could serve as dependent variables in a
regression equation, I decided to weight the scores in the
j problematic scales with scores from other dependency
scales which were not subject to the same distribution—or
which suffered from opposite problems. For instance, if a
scale had a spike on the highest value (i.e., 5) and j
■ another scale had a spike on the lowest value (i.e., 1), I |
|
i would weight the raw scores on one with the scores on the
i
' other. The effect of this was to reduce (but not
eliminate) the spike on both to a reasonable level.10
10After trying several weights, I decided to use 85%
of the original variable and 15% of the weight variable in i
each case where such a manipulation was necessary. This ,
method contrasts to factor scores in that—although factor j
scores are also linear combinations of the original t
variable and others—the amount of the score which is !
comprised of the original and of the other variables is
not easy to determine, or immediately obvious. Since I j
chose the variables to use for weighting, I can account j
for the theoretical compromises made in the combination. j
These vary, as will be seen, from variable to variable, j
depending upon the combination necessary to achieve the !
desired distribution.
The result, of course, is that the variables created
through this method are actually linear combinations of
two otherwise distinct variables. In regression analyses,
the variance explained by the independent variables will
be partially due to the relationship between the
independent variable and the "blending” variable used to
reduce the spike. The regression analyses using these
weighted variables must be interpreted with particular
care. The following variables are computed in this
fashion and used when dependency intensity is a dependent
variable in regression analysis:
For Newspapers: (1) Social Understanding (combined
with newspaper self understanding), (2) Self Understanding
(combined with newspaper social understanding), (3)
Interaction Orientation (combined with newspaper action
orientation), (4) Action Orientation (combined with
newspaper interaction orientation).
For Radio: (1) Self Understanding (combined with
newspaper action orientation), (2) Interaction Orientation
(combined with newspaper action orientation), (3) Social
Play (combined with newspaper action orientation).
For Magazines: (1) Interaction Orientation (combined
with newspaper action orientation), (2) Social Play
(combined with newspaper action orientation).
133
For Television: (1) Social Understanding (combined
with television interaction orientation), (2) Interaction
Orientation (combined with television social
understanding), (3) Solitary Play (combined with
television interaction orientation).
In order for the manipulation described above to be
successful, two criteria had to be met by the two
variables. (1) Their distributions had to be sufficiently
different that the combination of the two would
significantly change the distribution of the target
variable, and (2) they had to be at least moderately
correlated, so that high values on one were associated
1
with high values on the other (i.e. if a subject was on
the spike on one variable, he or she was likely to be on
I
the tail of the other variable). The reason that the J
scale for newspaper action orientation is used as the j
weighting scale so often above is that it was one of two '
i ;
j with a spike on the high end of the scale, and of the two j
was more often correlated with the target variable.
j One alternative to the weighting method described j
| above would have been the use of factor scores, which are
| standardized, rather than summative scale scores as
I j
i measures of the intensity of dependency relations. The
scale scores were preferred for two reasons: (1) Scale
scores are more readily understood as direct j
representations of the responses rather than the results
of statistical manipulation through factor analysis, and
(2) Factor scores, like the weighted scale scores
described above, are linear combinations of items in the
overall (18-item) scale. But unlike the weightings
described above, it is difficult to specifically account
for the amount of the magnitude in a factor score which is
attributable to items not designed to measure the concept
the score represents. Using the weighted scores allows at
least a modest amount of awareness of the sources of
potential confounds. I did compute some of the regression
results discussed below using factor scores, and did not
i
discover any appreciable differences in the thrust of the j
results. I remain confident that the weighted scale i
I
scores provide acceptable measures of the concepts they I
i
represent for present purposes.
Threat Perceptions and the Intensity and Substantive Scope
of Dependency Relations
Multiple regression analysis was conducted to test j
i
the fit of the data to the model of a parabola or bell
shaped curve when threat perceptions were the independent j
i
and the intensity of dependency relations the dependent |
j
variables. Oneway analyses of variance tested the j
substantive scope variables in their relation to threat j
I
perceptions. The results will be discussed for the ;
135
Table 5.12 Regression analysis of threat perceptions and media system dependency relations
for newspapers.
Combined sample
Independent Variables
17
Threat^
9
Threat
12
squared
9
Substantive Motivations B13 Intercept R B Intercept R n
Social understanding .133** 3.82*** .01 .037 3.53*** .01 934
Self understanding .315*** 2.85*** .04 - .055 3.28*** .04 934
Interaction orientation .312*** 2. 66*** .04 - .048 3.04*** .04 935
Action orientation .236*** 3.64*** .03 - .072 4.21*** .03 935
Austin
Independent Variables
Threat Threat squared
9
Substantive Motivations B Intercept R2 B Intercept R2 n
Social understanding .150** 3.86*** .01 .084 3.20*** .01 569
Self understanding .367*** 2.90*** .06 - .059 3.37*** .06 569
Interaction orientation .333*** 2.67*** .05 - .075 3.26*** .05 571
Action orientation .234*** 3.71*** .03 - .031 3.95*** .03 571
San Bernardino
Independent Variables
Threat
9
Threat squared
Substantive Motivations B Intercept R2 B Intercept R2 n
Social understanding .103 3.75*** .00 - .057 4.20*** .00 364
Self understanding .231** 2.77*** .02
’ - “ Sos
3.29*** .02 364
Interaction orientation .277*** 2.64*** .03 -8.62 03 2.71*** .03 363
Action orientation .234** 3.53*** .02 - .152 4.72*** .02 363
* -
p < .05, **
= P
< . 01, *** = p < .001
j 11The coefficients reported for this variable are j
(those calculated without controlling for the squared i
values of threat. 1
I
i
12This variable represents the non-linear association
between threat perceptions and dependency relations. The
regression coefficients reported for this variable are
calculated controlling for the linear association between ;
threat perceptions and dependency relations. j
I
13Unstandardized regression coefficient j
13 6^
combined data set and the separate Austin and San
Bernardino samples for each medium.
Newspapers
The regression analyses for newspapers showed no
I
support for the hypothesis that threat perceptions were
related to dependency relations in a non-linear fashion,
and highly significant support for a linear relationship
i
j between the two (see Table 5.12). In neither the combined
idata set, the Austin data, nor the San Bernardino data did
1 the solution offered by the parabolic equation generally
i
I explain significantly more variance than a linear
J equation. Higher perceptions of threat were associated
t
I with higher intensity dependency relations in the social j
understanding, self understanding, interaction |
i , 1
i orientation, and action orientation substantive j
] I
J motivations.14 The sole exception to this pattern was in
San Bernardino's social understanding analysis, where
neither the linear nor the non-linear equations J
demonstrated a significant relationship between threat and
the intensity of this dependency relation.
In the analyses of variance, a different pattern is
seen. A wide substantive scope—i.e. intense dependency j
relations with newspapers in two or more substantive
14Each of these dependent variables was weighted to
■ produce a more normal distribution, as explained above.
i ___________ _____________________________ ______________
137
Table 5.13 Analysis of variance between threat
perceptions and the substantive scope of
media system dependency relations for
newspapers. Scheffe test.
Threat perceptions by Substantive scope
Combined sample
Group
1 2 3
I N Mean
! 1. Low scope 252 -.14
! 2. Moderate scope 519 -.02
| 3. High scope 266 .17
* *
i
i
Threat perceptions by Substantive scope
Austin
Group
N Mean
1. Low scope 147 -.19
2. Moderate scope 310 -.01
*
3. High scope 162 .18
* *
Threat perceptions by Substantive scope
San Bernardino
Group
1 2 3
N Mean
1. Low scope 105 -.08
2. Moderate scope 209 -.05
3. High scope 104 .15
* *
I
I
I * = p < .05
138
motivations—was associated with significantly higher mean
i
scores on the threat perception scale (see Table 5.13).
Low threat perceptions were associated with the lowest
substantive scope, while the middle range of mean threat
j
perceptions were reported by those in the moderate range j
i
of substantive scope. The differences in threat J
j
1 perceptions between the middle and low substantive scope '
i '
j groups were not statistically significant, but the highest i
I mean threat score, for the highest substantive scope i
\ J
| group, was significantly different from both other groups ;
j j
j (p < .05).This describes a linear progression. This is |
]
1 contrary to Hlc, which predicts an inverted U, or j
i 1
! parabola. This pattern is seen in all data sets (in \
I ;
i Austin, all difference between groups are statistically j
I
l j
| significant). 1
i
Radio j
In the combined data set, the results of the
I
(
regression equation for radio are similar to those for
I
newspapers (see Table 5.14).15 The linear equation !
provides a highly significant fit to the data in all j
substantive motivations, and the parabolic equation does '
i
I
not significantly improve upon this fit. In the Austin j
data set, this is not the case for the self understanding j
15For radio, the self understanding and interaction
orientation variables were weighted to produce a more
normal distribution.
| ___________________________________________________
139
Table 5.14 Regression analysis of threat perceptions and media system dependency relations
for radio.
Combined sample
Independent Variables
Substantive Motivations B18
Threat^8
Intercept R2
Threat
B
squared^7
Intercept R2 n
Social understanding .237*** 3.16*** .02 .050 2.77*** .02 964
Self understanding .244*** 2.41*** .03 - .046 2.77*** .03 917
Interaction orientation .192*** 2.39*** .02
' ’ °2-03
9.89 03
2.58*** .02 915
Action orientation .275*** 2.64*** .03 2.56*** .03 940
Substantive Motivations B
Austin
Independent Variables
Threat Threat
Intercept R B
squared
Intercept R2 n
Social understanding .253*** 3.09*** .02 - .057 3.54*** .02 582
Self understanding .276*** 2.38*** .05 - .126* 3.37*** .05 562
Interaction orientation .243*** 2.34*** .04 - . 121* 3.30*** .04 564
Action orientation .294*** 2.64*** .04 - .012 2.73*** .03 576
Substantive Motivations B
San Bernardino
Independent Variables
Threat _ Threat
Intercept R B
squared
Intercept R2 n
Social understanding .218** 3.27*** .01 .228* 1.47 .03 381
Self understanding .197** 2.45*** .02 .084 1.79** .02 354
Interaction orientation .117 2.47*** .00 .132 1.43* .01 350
Action orientation .245** 2.64*** .02 .041 2.31** .02 363
* -
p < .05, **
= P
< . 01, *** = p < .001
16The coefficients reported for this variable are
those calculated without controlling for the squared
'values of threat.
I |
17This variable represents the non-linear association j
between threat perceptions and dependency relations. The
regression coefficients reported for this variable are j
calculated controlling for the linear association between
j threat perceptions and dependency relations.
i
| 18Unstandardized regression coefficient
140
and the interaction orientation motivations. In both
cases, the non-linear equation is a significantly (p <
.03) better fit to the data than the linear equation. In
San Bernardino, the solution for the social understanding
motivation is best represented by a non-linear equation,
but the value for B, is positive. thereby describing a U-
shaped curve, contrary to HI. Also in San Bernardino
neither the linear nor non-linear equation provides a
significant fit to the data in the interaction orientation
motivation. It would appear that radio differs from
newspapers in its relation to threat perceptions, and that
radio in the two communities activates different
dependency relations when threat perceptions are
controlled.
In the substantive scope analyses of variance,
results similar to those for newspapers are found (see
Table 5.15). In the combined and Austin data sets, threat
perceptions rise as substantive scope widens, and all
differences between groups are statistically significant
(p < .05). The overall pattern is linear, as noted for
newspapers. In San Bernardino the linear pattern is
observed, but the only statistically significant
difference is between the low and high scope groups—the
moderate scope group is too close to both extremes to be
significantly different.
141
Table 5.15 Analysis of variance between threat
perceptions and the substantive scope of
media system dependency relations for
radio. Scheffe test.
Threat perceptions by Substantive scope
Combined sample
Group
1 2 3
N Mean
1. Low scope 269 -.17
2. Moderate scope 511 -.00
*
3. High scope 257 . 17
* *
Threat perceptions by Substantive scope
Austin
Group
1 2 3
N Mean
1. Low scope 176 -.18
2. Moderate scope 297 .01
*
3. High scope 146 .19
* *
Threat perceptions by Substantive scope
San Bernardino
Group
1
N Mean
1. Low scope 93 -.16
2. Moderate scope 214 -.02
3. High scope 111 .14
*
142
Table 5.16 Regression analysis of threat perceptions and media system dependency relations
for magazines.
Combined sample
Independent Variables
Threat19
• y
Threat squared20
Substantive Motivations B21 Intercept R2 B Intercept R2 n
Social understanding .128** 2.67*** .01 .086 1.99*** .01 947
Self understanding .246*** 2.98*** .02 - .178** 4.38*** .03 935
Interaction orientation .207*** 3.00*** .02 - .108 3.85*** .02 918
Action orientation .227*** 2.44*** .02 - .039 2.74*** .02 935
Austin
I independent Variables
Threat Threat squared
• y
Substantive Motivations B Intercept R2 B Intercept R2 n
Social understanding .142* 2.73*** .01 3.29'03 2.71*** .01 574
Self understanding .234*** 3.01*** .02 - .166* 4.32*** .03 569
Interaction orientation .214** 3.01*** .02 - .105 3.84*** .02 562
Action orientation .180** 2.43*** .01 - .093 3.16*** .02 571
San Bernardino
I independent Variables
Threat
• y
Threat squared
Substantive Motivations B Intercept R2 B Intercept R2 n
Social inderstanding .100 2.56*** .00 .235* .71 .02 372
Self understanding .263** 2.93*** .02 - .206 4.56*** .03 365
Interaction orientation .195* 2.97*** .01 - .120 3.92*** .01 354
Action orientation .302*** 2.46*** .04 .075 1. 86* .04 363
* s
p < .05. **
= P
< . 01, *** = p < .001
19The coefficients reported for this variable are
those calculated without controlling for the squared
values of threat.
20This variable represents the non-linear association
between threat perceptions and dependency relations. The
regression coefficients reported for this variable are
calculated controlling for the linear association between
threat perceptions and dependency relations.
21Unstandardized regression coefficient
143
Magazines
The combined data set provides support for the non
linear relationship between threat perceptions and
dependency relations for the self understanding motivation
for magazines (see Table 5.16).22 This is the first
instance yet reported where the combined data set provides
support for a non-linear relationship (all linear
solutions are significant). In Austin, as in the combined
set, the self understanding motivation provides support
for a non-linear fit to the data in the hypothesized form.
All other motivations remain better explained by a linear
equation.
In San Bernardino, the linear equation does not
provide a significant fit to the data for social
understanding, but the improvement offered by the non
linear equation provides a significant fit (p = .016).
The non-linear equation for social understanding in San
Bernardino does not describe a parabola, however, but an
"S-curve," with higher threat perceptions related to
rapidly ascending intensity scores for this dependency
relation. Non-linear equations provide better fits for no
other motivation in the San Bernardino data, and linear
solutions are significant in all other motivations.
22For magazines, the interaction orientation variable
was weighted to produce a more normal distribution.
144
Table 5.17 Analysis of variance between threat
perceptions and the substantive scope of
media system dependency relations for
magazines. Scheffe test.
Threat perceptions by Substantive scope
Combined sample
Group
1 2 3
N Mean
1. Low scope 237 -.10
2. Moderate scope 549 -.02
3. High scope 251 . 13
* *
Threat perceptions by Substantive scope
Austin
Group
1 2 3
N Mean
1. Low scope 140 -.12
2. Moderate scope 327 . 00
3. High scope 152 . 10
*
(No significant differences)
Threat perceptions by Substantive scope
San Bernardino
Group
1 2 3
N Mean
1. Low scope 97 -.07
2 . Moderate scope 222 -.05
3. High scope 99 .16
* = p < .05
145
Table 5.18 Regression analysis of threat perceptions and media system dependency relations
for television.
Combined sample
Independent Variables
9E
Threat23 _ Threat
J/
squared
5
Substantive Motivations B25 Intercept R B Intercept R2 n
Social understanding .288*** 3.92*** .05 - .067 4.45*** .05 932
Self understanding .413*** 3.11*** .06 - .065 3.62*** .06 950
Interaction orientation .383*** 2.96*** .06 - .031 3.20*** .06 932
Action orientation .442*** 2. 88*** .07 - .041 2.56*** .07 947
Austin
Independent Variables
Threat _ Threat squared
0
Substantive Motivations B Intercept R B Intercept R2 n
Social understanding .247*** 3.98*** .04 - .081 4.62*** .04 569
Self understanding .421*** 3.12*** .06 - .092 3.84*** .06 575
Interaction orientation .393*** 2.98*** .07 - .082 3.62*** .07 569
Action orientation .393*** 2.92*** .06 - .024 3.10*** .06 576
San Bernardino
Independent Variables
Threat _ Threat squared
Substantive Motivations B Intercept R B Intercept R2 n
Social understanding .348*** 3.83*** .06 - .028 4.05*** .06 362
Self understanding .400*** 3.10*** .05 - .017 3.24*** .05 374
Interaction orientation .366*** 2.93*** .05 .064 2.43** .05 362
Action orientation .515*** 2.83*** .08 .177 1.44 .09 370
* -
p < .05. ** = p < . 01, *** = p < .001
23The coefficients reported for this variable are
those calculated without controlling for the squared
values of threat.
24This variable represents the non-linear association
between threat perceptions and dependency relations. The
regression coefficients reported for this variable are
calculated controlling for the linear association between
threat perceptions and dependency relations.
25Unstandardized regression coefficient
146
Table 5.19 Analysis of variance between threat
perceptions and the substantive scope of
media system dependency relations for
television. Scheffe test.
Threat perceptions by Substantive scope
Combined sample
Group
1 2 3
N Mean
1. Low scope 173 -.07
2. Moderate scope 533 -.11
3. High scope 331 .20
• k k
Threat perceptions by Substantive scope
Austin
Group
1 2 3
N Mean
1. Low scope 105 -.02
2. Moderate scope 327 -.12
3. High scope 187 .21
* *
Threat perceptions by Substantive scope
San Bernardino
Group
1 2
N Mean
1. Low scope 68 -.16
2. Moderate scope 206 -.09
3. High scope 144 . 19
* *
* = p < .05
147
The substantive scope results in the combined and San
Bernardino data sets show a by-now familiar pattern (see
Table 5.17). Threat perceptions rise as substantive scope
widens, although the difference between the low and
moderate scope groups is not statistically significant.
In the Austin data, only the difference between the low
and high scope groups is significant.
Television
The combined data set provides no support for the
hypothesis of non-linear relations between threat
perceptions and the intensity of dependency relations (see
Table 5.18).26 Linear solutions are significant in all
motivations. In Austin and in San Bernardino the same is
true. These results are most similar to those reported
for newspapers.
In the analyses of variance, the substantive scope
hypothesis is again given no support, and the results are
extremely consistent with those noted above for
newspapers, radio, and magazines (see Table 5.19). Once
again, the differences between the threat perceptions of
those in all scope groups are significant, and threat
perceptions rise as scope widens. In San Bernardino, the
26For television, the social understanding and
interaction orientation variables were weighted to produce
a more normal distribution.
148
difference between the low and moderate scope groups is
not significant.
Overall, these analyses suggest that there are
occasionally significant differences between respondents
in the two cities in the relationships they report between
threat perceptions and media system dependency relations,
and important differences as well between media in the
sorts of relations people report with them. Generally,
linear relations between threat perceptions and the
intensity of dependency relations are found, contrary to
my main hypothesis.
The substantive scope results are so consistent that
one can hardly help but conclude that substantive scope is
related to threat in a manner almost directly opposite to
that hypothesized—i.e. it appears that threat perceptions
rise as substantive scope widens.
Threat Perceptions and the Media Scope of Dependency
Relations
The relationship between media scope and dependency
relations was examined through analyses of variance.
These analyses are performed separately for each
substantive motivation. People are placed into groups of
low, medium, and high depending on how many media with
which they report intense dependency relations in each
substantive motivation. High intensity with three or four
149
Table 5.20 Analysis of variance between threat
perceptions and the media scope of media
system dependency relations for social
understanding. Scheffe test.
Threat perceptions by Media scope
Combined sample
Group
1 2 3
N Mean
1. Low scope 393 -.08
2. Moderate scope 475 -.01
3. High scope 169 . 18
* *
Threat perceptions by Media scope
Austin
Group
1
N Mean
1. Low scope 231 -.09
2. Moderate scope 298 .01
3. High scope 90 .19
*
Threat perceptions by Media scope
San Bernardino
Group
1 2 3
N Mean
1. Low scope 162 -.06
2. Moderate scope 177 -.04
3. High scope 79 . 16
*
Table 5.21 Analysis of variance between threat
perceptions and the media scope of media
system dependency relations for self
understanding. Scheffe test.
Threat perceptions by Media scope
Combined sample
Group
1 2 3
N Mean
1. Low scope 577 -.10
2. Moderate scope 292 .06
*
3. High scope 168 .22
* *
Threat perceptions by Media scope
Austin
Group
1 2
N Mean
1. Low scope 356 -.11
*
2. Moderate scope 171 . 10
3. High scope 92 .22
Threat perceptions by Media scope
San Bernardino
Group
1 2 3
N Mean
1. Low scope 221 -.10
2. Moderate scope 121 .01
3. High scope 76 .22 *
media places one in the "high” category, while high
intensity with one or two media places one in the "medium"
category, and high intensity in no media for a given
jsubstantive motivation places one in the "low" category.
1 Social Understanding
i
I Media scope for social understanding was related to
!threat perceptions such that the higher one's threat
jperceptions, the wider the media scope one reported (see
liable 5.20). The difference between the threat
perceptions of the lowest and middle scope groups was not
statistically significant, but the widest scope group's
jthreat perceptions were significantly higher than the
|other two groups'. This essential pattern was found in
;the combined data set as well as in the individual cities.
This result does not support hypothesis Id, which
predicted a distribution with low threat perceptions for
i
jthe lowest and widest media scope and the highest threat
perceptions associated with the highest media scope.
Self Understanding
i As with social understanding above, the essential
!
j result in this motivation is that the wider one's media
i
!scope, the higher his or her threat perception is likely
to be (see Table 5.21). This pattern is observed in all
three data sets. In the combined set the mean threat
scores of each scope group are significantly different
Table 5.22 Analysis of variance between threat
perceptions and the media scope of media
system dependency relations for
interaction orientation. Scheffe test.
Threat perceptions by Media scope
Combined sample
Group
1 2 3
N Mean
1. Low scope 562 -.08
* *
2. Moderate scope 276 .04
3. High scope 199 . 16
Threat perceptions by Media scope
Austin
Group
1 2 3
N Mean
1. Low scope 357 -.08
* *
2. Moderate scope 162 .08
3. High scope 100 .15
Threat perceptions by Media scope
San Bernardino
Group
1 2 3
N Mean
1. Low scope 205 -.08
2. Moderate scope 114 -.03
3. High scope 99 . 17
*
153
Table 5.23 Analysis of variance between threat
perceptions and the media scope of media
i system dependency relations for action
| orientation. Scheffe test.
}
I Threat perceptions by Media scope
j Combined sample
Group
1 2 3
N Mean
1. Low scope 514 -.13
* *
2. Moderate scope 335 .08
3. High scope 188 .19
Threat perceptions by Media scope
Austin
Group
1 2 3
N Mean
1. Low scope 318 -.11
* *
2. Moderate scope 200 .10
3. High scope 101 . 14
Threat perceptions by Media scope
San Bernardino
Group
N Mean
196 -.16
135 .05
87 . 24
1 2 3
II. Low scope 196 -.16 * *
2. Moderate scope
3. High scope
* =
p < .05
Table 5.24 Analysis of variance between threat
perceptions and the media scope of media
system dependency relations for social
play. Scheffe test.
Threat perceptions by Media scope
Combined sample
I
I
Group
1 2 3
N Mean
1. Low scope 520 -.12
* *
2. Moderate scope 350 .08
3. High scope 167 .16
Threat perceptions by Media scope
Austin
I
Group
1 2 3
N Mean
1. Low scope 334 -.10
* *
2. Moderate scope 203 .08
3. High scope 82 .19
Threat perceptions by Media scope
San Bernardino
Group
1 2 3
N Mean
1. Low scope 186 -.15
* *
2. Moderate scope 147 .09
3. High scope 85 .13
from one another. In Austin, the lowest scope group has a
score significantly lower than the other two groups, and
in San Bernardino the lowest group is significantly
different only from the highest scope group.
i
Interaction Orientation
This motivation's results emulate the two above (see
[Table 5.22). The same linear pattern of media scope and
jthreat perceptions is observed, and this holds in each
jcity as well. The statistically significant differences
l
1 in the combined data set and Austin are between the lowest
i
i
'scope group and the other two. In San Bernardino only the
mean threat perceptions of the lowest and highest scope
j
groups are significantly different.Action Orientation
j The by now familiar pattern which has been observed
j
jin three motivations is seen for action orientation as
l
well (see Table 5.23). In each data set, the only
statistically significant differences are between the
!
!lowest scope group and the other two groups. Mean threat
perceptions rise as the breadth of one's media scope
(
!widens.
|
|Social Play
j The relations between threat perceptions and media
i
j scope in this motivation directly mirror those in the
i
{action orientation motivation, including the pattern of
significance levels (see Table 5.24).
Table 5.25 Analysis of variance between threat
perceptions and the media scope of media
system dependency relations for solitary
play. Scheffe test.
Threat perceptions by Media scope
Combined sample
Group
1 2 3
1. Low scope
N
481
Mean
-.10
4c 4c
2. Moderate scope 372 .04
3. High scope 184 . 16
Threat perceptions by Media scope
1. Low scope
N
295
Austin
Group
Mean
-.09
1 2 3
2. Moderate scope 230 .03
3. High scope 94 .22
*
Threat perceptions by Media scope
San Bernardino
Group
1 2 3
N Mean
1. Low scope 186 -.12
* *
2. Moderate scope 142 .06
3. High scope 90 .11
j 157
iSolitary Plav
1 No change from the patterns observed in the other
1
ifive substantive motivations occurs for solitary play (see
I
jTable 5.25). Higher mean threat perceptions are
(associated with a wider media scope.
In all, the media scope hypothesis (Hid) is offered
no support in any substantive motivation in any of the
data sets. One may reliably claim support for an
I
j alternative hypothesis—that high threat perceptions are
|associated with a wider media scope. This alternative
I hypothesis is supported in every instance, with varying
|patterns of significant differences between the means of
i
the three groups.
Threat Perceptions, Education, Income, and Age
A multiple regression equation was calculated for
i
each sample with the demographic variables and the
intensity of dependency relation as independent variables
Ipredicting the level of threat perceptions.27 Hypothesis
two predicted significant negative correlations between
education and threat and between income and threat when
!other demographic and dependency variables were
controlled. In the combined data set, the partial
! 27As independent variables, the suramative dependency
i scales rather than the weighted scales are suitable for
(all substantive motivations. The independent variables in
(a regression equation need not be normally distributed.
158
correlation between education and threat perceptions was,
as expected, negative, but it was not significant (p =
.206). Income, however, was negatively and highly
significantly correlated with threat perceptions (r =
-.15, p = .0001). Hypothesis two is only partially
supported by these results. The intercorrelation between
income and education may have diluted the magnitude, and
therefore the significance, of the latter's correlation in
this equation. The same equation calculated separately
for each city yields the same essential results.
Age was hypothesized to be significantly positively
correlated with threat perceptions when the other
demographic and the dependency variables were controlled
in the multiple regression equation. In the combined
sample, age was not correlated with threat perceptions to
any remarkable degree (r = .01; p = .7323). In the city
samples, this result is partly explained. In Austin, the
magnitude of the positive partial correlation is higher
than in the combined sample (r = .05), and while still
insignificant, it is much closer to significance than in
the combined sample (p = .3038). In San Bernardino,
however, the partial correlation between age and threat
perceptions is negative (r = -.05), and still not
significant (p = .4477). Clearly, the combination of
these relationships would produce the result seen in the
I combined data set. Hypothesis three receives no support,
i
i
ihowever, from any of these equations.
j Other significant relationships in these equations
!are worth noting. Overall, it is interesting to note that
>
■in the combined data set only 16.7% of the variance in
I
'threat perceptions is explained by all these variables
i
icombined. Of course, one key variable in threat
'perceptions not included in the analysis is the actual
i probability of these people suffering harm from one or
i
I
Imore of the threats referred to in the scale. Presumably
j a great deal of variance in their perceptions is explained
j
jby that missing variable. Nonetheless, for such powerful
|demographic variables such as age, income, education,
I ethnicity, sex, and household population (children and
I
|adults) to account for so little variance is somewhat
t
surprising. In the combined data set, the significant
i
[variables besides income include the number of children in
ithe household (B — .06, p = .0035), sex (B = -.17,28
i p = .0004), ethnicity (B = -.26,29 p < .0001), and the
[intensity of the newspaper self understanding (B = .09,
28A negative slope for this dummy variable indicates
that women perceive more threat than men.
29A negative slope on this dummy variable indicates
that those identifying themselves with a non-white ethnic
group perceive more threat than whites.
p = .025) and radio interaction orientation (B = -.13,
p = .0134) dependency relations.
In the Austin equation, the same general pattern of
significant relations as in the combined set was observed
i
(except that radio interaction orientation was not
significantly related to threat perceptions) and 18% of
the variance in threat perceptions was explained. In San
Bernardino, however, 13.9% of the variance is explained
I
i
!and the number of children in the household is not a
i
!significant factor, and ethnicity is only moderately
jsignificant (B = -.18, p = .05). The radio interaction
I
jorientation (B = -.18, p = .021) and the television social
understanding (B = .10, p = .0517) dependency relations
;are the only significant dependency variables in the San
|Bernardino equation.
; Of the three demographic variables about which
hypotheses were made—education, income, and age—only
[income proved to be related to threat perceptions in the
'manner hypothesized. This was true in the combined set as
jwell as in each individual city. Dependency relations
were not often significant predictors of threat
perceptions, and never as strong as threat perceptions
were as predictors of dependency relations in the analyses
i
;reported above.
I 161
j Threat perceptions are highest, it appears, among
|those most likely to be threatened by social conditions in
the contemporary American environment—the poorer, non
white, female population with children. To test the
ability of the threat scale to capture differences in the
specific threats to different parts of the population, a
second set of regression equations was performed, with
ethnicity, sex, age, and income used to predict responses
to each separate item on the scale. There are eight items
in the threat scale, and each refers to a distinct type of
|hazard. Of the eight regression equations, in only two
I was age a significant predictor of responses—for the item
i
regarding the rising cost of medical care and for the item
jregarding loss of one's job. These two items are
!certainly likely to strike a chord with older respondents,
i
for whom each of these concerns are likely to be more
weighty than for the younger respondents (these hazards
are not unrelated, since employers provide most medical
1 insurance). The only item in which income was not a
j
jsignificant predictor was the threat of being mugged or
robbed. The only item in which sex was not a significant
i
»
predictor was fear of losing one's job. Ethnic minorities
perceive more threat from all sources.
162
; Media Use, Threat, and Dependency Relations
i
j Multiple regression equations were computed using the
|measures of time spent using the four mass media as the
dependent variables and threat perceptions, dependency j
relations, and the demographic variables as the I
independent variables. A second equation was then
|computed adding the use of other media to the variables (
jpreviously included. In this way, the effects of the use |
1 I
t ‘
I ,
of other media on the use of one medium could be i
controlled for. The results of these analyses will be '
presented separately for each medium.
J
|
Use of the print media—newspapers and magazines—was
not distributed normally among the sample. Unlike the
!
|problematic dependency variables, however, the
i
|distribution of readership of newspapers and magazines was
r
[such that most people reported no reading or very little
reading, and a straightforward logarithmic
I transformation—in which the natural log of each variable
j \
jwas used to represent the variable itself—provided a
I dependent variable with a normal distribution suitable for
sregression analysis. In the analyses discussed below,
i
I
I when newspaper or magazine use is discussed it is the
i
jtransformed variables which were used in the regression
'equation. When readership is an independent variable, no
such transformation was performed.
j Newspapers
The amount of variance in newspaper use explained by
variables other than use of other media was 22.3% in the
combined sample, 21.9% in Austin, and 21.8% in San
i
Bernardino. Threat perceptions were not significantly
related to reports of daily newspaper use (i.e. hours read
each day) in any of the three samples in this equation.
Significant newspaper dependency variables in the combined
'sample and in the Austin sample include social
understanding and solitary play. In San Bernardino, only
social understanding reaches significance,
j Demographic variables associated with readership in
|the combined sample include sex (men read more), age (the
I
[
;older read more), income (lower income is associated with
I
more time spent reading), and the number of adults in the
household (more adults associated with more reading). In
:Austin, only income and age were significantly related to
|readership, and in the same manner as in the combined
!sample. In San Bernardino men read significantly more
I
than women, and the unemployed read more than the full or
part time employed.30
30This dummy variable made no distinction between the
unemployed who were not looking for work and those who
were.
Table 5.2 6 Regression analysis of newspaper use and
threat perceptions, media system dependency
relations, demographic variables, and
use of other media.
Combined sample
Independent Variables Newspaper use
Partial
B Correlation
Threat - 5.17'03 -.005
Social understanding .197*** .197
Self understanding .013 .012
Interaction orientation .030 .028
Action orientation - .024 -.026
Social play .060 . 074
Solitary play .161*** . 205
Adults in household .051 .072
Sex31 .137** . 104
Years lived in community 1.15*03 .021
Children in household - .030 -.047
Race32 - 8.49’03 -.001
Employment status33 .039 -.024
Education
1
t f i
•
to
o
1
o
-.002
Income .021 -.068
Age
9,84-05***
. 170
Television use 4.62'04* . 090
Magazine use 8.16'04*** .238
Radio use 3.20‘04* .084
Intercept = 6.87*** R2 = .28 N = 706
* = p < .05 ** = p < .01 *** = p < .
001
i
I
!
I
I
31 Dummy coded such that 0 = female, 1 = male.
32Dummy coded such that 0 = non-white, 1 = white.
i 33Dummy coded such that 0 = not employed, 1 = employed
at least part time.
When the use of other media was added to the
equation, the pattern of significant relationships changed
i
{somewhat (see Table 5.26). The equation in the combined
i
isample explained 28.4% of the variance, 29.2% in Austin,
and 25.8% in San Bernardino. In the combined sample, the
more time one reported using radio, television, and
magazines the more time he or she was likely to report
I
heading newspapers. In Austin, this was not true of radio
i
■ use and in San Bernardino it was not true of television
J
juse. Controlling for other media use did not make threat
■perceptions any more significant as predictors of
newspaper reading in either city—in fact they greatly
reduced the probability of significance.
i
! In the combined sample, the intensity of social
r
understanding and solitary play dependency relations
iremained highly significantly and positively associated
i
with readership, and the intensity of the social play
motivation approached moderate significance as well when
other use variables were controlled (p = .0535). The
pattern of significant relations between dependency
relations and newspaper reading in Austin did not change
iin this equation. In San Bernardino, social understanding
remained significant, but solitary play also achieved
!significance after controlling for other use variables.
The demographic variables which had been associated
with newspaper reading in the combined sample in the first
equation (sex, age, income, and the number of adults in
the household) did not all remain significant. The number
of adults in the household fell to a marginally
significant probability (p = .0597), income fell further
into insignificance (p = .0757), while age and sex
remained highly significant predictors (with men still
reading more than women). In Austin, where income and age
had been the only significant demographic predictors of
readership in the first equation, there was no change in
(the significant demographic variables. In San Bernardino,
isex and employment status had been significant predictors,
I
jbut employment status did not remain significant after the
control for other media use.
Radio
j The initial equation, not controlling for time spent
[with other media, explained 19.5% of the variance in radio
listening in the combined sample, 18.8% in Austin, and
'18.8% in San Bernardino. Threat perceptions were not
i
!significantly associated with radio use in any sample. Of
^the dependency intensity variables in the combined sample,
|solitary play, interaction orientation, social
understanding, and action orientation all provided
significant predictors of radio listening. In Austin,
167
Table 5.27 Regression analysis of radio use and threat
perceptions, media system dependency
relations, demographic variables, and
use of other media.
Combined sample
jIndependent Variables Radio use
1
!
B
Partial
Correlation
i
1 Threat .763 .003
|Social understanding 27.891*** . 130
;Self understanding 3.241 .011
jInteraction orientation -23.778* -.075
1 Action orientation 16.820 .066
!Social play 4.012 .017
i Solitary play 40.075*** .214
Adults in household .256 .001
Sex34 - .763 -.002
'Years lived in community .726 .053
Children in household 7.222 .046
Race35 14.075 .036
‘ Employment status36 35.217* . 086
jEducation - 1.271 -.021
1 Income - 8.544** -.109
- .507 -.033
1 Television use .086 . 070
|Magazine use .038 . 045
;Newspaper use .144 .042
'intercept = -80.84 R2 = .20 N = 732
i * = p < .05 ** = p < .01
*** = p <
.001
I
<
*
I
I
\
34Dummy coded such that 0 = female, 1 = male.
| 35Dummy coded such that 0 = non-white, 1 = white.
' 36Dummy coded such that 0 = not employed, 1 = employed
iat least part time.
only solitary play and social understanding were j
significant predictors in this equation; in San
i
Bernardino, only solitary play. i
In the combined sample, the demographic variables j
which best predicted radio use when other media use was
^not controlled for were employment status (the employed
1
1
i
listened more) and income (the less affluent listened j
more). In Austin only employment status was significant, ^
and in San Bernardino only income (both in the same manner
[as in the combined sample).
i When use of other media was controlled in the second
i ‘
i '
[equation, 20.1% of the variance was explained in the
combined sample, 19.2% in Austin, and 18.8% in San
Bernardino (see Table 5.27). The small increase in
explained variance makes it unsurprising that use of other
|media is not significant as a predictor of radio listening
jin any sample (although in Austin and the combined sample
television viewing approaches significance). Threat
perceptions remain virtually insignificant in their
i
[relation to media use. The only change in dependency
!relations is that action orientation is no longer
jsignificant in the combined sample (the Austin and San
Bernardino samples showed no changes in this regard). No
change in the demographic variables in any sample was
J
‘observed.
Table 5.28 Regression analysis of magazine use and
threat perceptions, media system dependency
relations, demographic variables, and
use of other media.
Combined sample
Independent Variables Magazine use
Partial
B Correlation
1
Threat .046 .035
Social understanding .061 .055
Self understanding .013 -.010
Interaction orientation .041 .034
Action orientation .055 -.044
!Social play 4 .97"03 .004
Solitary play .410*** .367
Adults in household .036 -.040
: Sex37
• t i
.157 .088
Years lived in community 3.79'03 .053
Children in household .015 -.018
Race38 - .140 -.069
Employment status39 .063 . 030
Education 2.93 03 .010
Income 4.75*03 .012
Age 8.62'03** .115
Television use 4.51'04 .073
Radio use 2.95'04 . 062
Newspaper use
3 # 48-03***
.201
Intercept = 2. 63*** R2 = .28 N = 684
* = p < .05 ** — p < .01
*** = p <
.001
37Dummy coded such that 0 = female, 1 = male.
^Dummy coded such that 0 = non-white, 1 = white.
39Dummy coded such that 0 = not employed, 1 = employed
at least part time.
Magazines
When other media use is not controlled, 23.8% of the
variance in magazine reading in the combined sample, 24%
in Austin, and 22.1% in San Bernardino is explained by
threat perceptions, the intensity of media dependency
relations, and demographic variables. Threat perceptions,
thowever, are not significant predictors of magazine use in
i
tany sample. In all samples, the intensity of the solitary
'play dependency relation is highly significant, but no
!other substantive motivation reaches significance.
Significant demographic variables in the combined sample
include sex (men read more), ethnicity (non-whites read
I
'more), and age (the older read more). In Austin, sex was
jnot a significant predictor, but ethnicity and age were
associated as in the combined sample. In San Bernardino,
t
t
only sex was significant (men read more).
| Controlling for other media use explains 27.6% of the
!
jvariance in readership in the combined sample, 28.2% in
I
^Austin, and 27.4% in San Bernardino (see Table 5.28). In
ithe combined sample and in Austin, time spent reading the
jnewspaper is the only significant use measure related to
i
imagazine reading. In San Bernardino, however, only radio
juse is significantly related to magazine reading. Threat
remains insignificant as a predictor in this equation in
each sample. No change in the significance of dependency
171
Table 5.29 Regression analysis of television use and
threat perceptions, media system dependency
relations, demographic variables, and use
of other media.
Combined sample
Independent Variables Television use
1
i
1
Partial
1
I B Correlation
Threat
_
4.642 -.022
Social understanding 27.066*** .137
Self understanding
—
.243 -.001
Interaction orientation 3.600 .018
Action orientation 1.232 .007
Social play
-
8.531 -.044
Solitary play 34.230*** .187
Adults in household 1.407 .010
Sex40 .400 . 002
Years lived in community .573 .052
Children in household 2.481 .020
Race41
—
18.293 -.052
Employment status42
—
68.675*** -.213
Education
-
4.654 -.097
Income
-
6.919** -.111
Age .409 .036
Magazine use .062* .090
Radio use .050 .066
Newspaper use .225* .082
Intercept =56.46 R = .22 N = 750
*=p<.05 **=p<. 01
*** = p <
.001
! 40Dummy coded such that 0 = female, 1 = male.
j 41 Dummy coded such that 0 = non-white, 1 = white.
42Dummy coded such that 0 = not employed, 1 = employed
iat least part time.
172!
relations was observed—solitary play remains highly
i
significant, and the only substantive motivation which is j
i
significant. Among the demographic variables the only
change was that ethnicity became insignificant in the
i
combined sample and Austin when other media use was
I
controlled. j
Television
Without controlling for the use of other media, the t
combined sample explained 20.5% of the variance in
television use, 20.6% in Austin, and 20% in San Bernardino
I
I '
through threat perceptions, dependency relations, and
demographic variables. Threat perceptions failed to show
a significant relationship to television use. In the
combined sample and Austin, dependency relations of
solitary play and social understanding were significant
;predictors of television use. In San Bernardino social
i
!play approached marginal significance (p = .0578) and
fsocial understanding was highly significant,
j Among the demographic variables, employment status
i (the unemployed watch more), income (the less affluent
1
;watch more), and education (the less well-educated watch
|more) were all significantly related to television use in
I
|the combined sample and in Austin. In San Bernardino,
t
only employment status proved significantly related.
! : I
j 173:
i
When use of the other media was controlled, 21.9% of i
i
\
the variance in television viewing in the combined sample, j
21.9% in Austin, and 22.5% in San Bernardino was explained
(see Table 5.29). In the combined sample, time spent t
I
i
reading newspapers and time spent reading magazines were
significant predictors of time spent with television. In
Austin, only newspaper use was significant, and in San '
Bernardino only magazine use. As in each of the other !
!regression equations for media use, there was no
I
i
significant linear relation between threat perceptions and
[television use. There was no change in the pattern of
i
,significant relations between either dependency or
I
i
jdemographic variables when other media use was controlled.
i
j In summary, significantly more variance in the amount
of time spent using the print media is explained when use
;of other media—including broadcast media—are controlled.
!In the dependency motivations, even with use of other
media controlled the intensity of the solitary play
dependency relation remains highly significant as a
I
I predictor of use of most media, and approaches
significance whenever it is not plainly significant.
;Social understanding is the only other motivation which
|consistently reaches significance as a predictor of use.
Income, age, education, and employment status were all
jobserved to be significant predictors of media use for
I some media. For present purposes, the most striking
J finding is the complete absence of any significant linear
I relationship between threat perceptions and media use for
*
j any medium.
i Summary
Investigation of the first research question showed
that the highest threat perceptions are associated with
[
j the members of society most vulnerable to social problems—
i
minority groups, women, the less affluent, and the less
educated. This pattern was discernable in both cities
studied. The second research question, concerning the
j dimensions of dependency in the data collected in Austin
i
; and San Bernardino, provided evidence that between five
and seven empirical factors were present, and provided
enough support for use of summative scale scores in
| subsequent analyses. Some persistent problems with the
j dependency scale were noted, such as failure to
consistently distinguish between self understanding and
i interaction orientation and weak measurement of the social
i
play dimension.
The first hypotheses, concerning the nature of the
i relationship between perceptions of threat and the
i
i
intensity and scope of media system dependency relations
, were generally not supported. Support for hypotheses la
17 51
I
and lb—i.e. those hypotheses regarding intensity—was I
provided in the following analyses:
; Radio:
i Austin—Self understanding and interaction j
\ orientation ;
I '
! . i
j Magazines: |
j Combined sample—Self understanding
J Austin—Self understanding
! In San Bernardino, no significant relationship |
| |
|between the intensity of newspaper social understanding >
|dependency and threat perceptions or between the intensity
1 of radio interaction orientation dependency and threat
i
|perceptions was observed. In San Bernardino, a non-linear
jregression equation was significant for magazines, but the
I
!equation describes an "S-curve," not the hypothesized
1 parabola. In all other cases in the combined sample,
'Austin sample, and San Bernardino sample a significant
’linear relationship between the intensity of dependency
j relations and threat perceptions was observed.
!
; The test of hypothesis lc showed no support at all
for the parabolic relationship between the substantive
!scope of media dependency relations and threat
!
perceptions. Statistically significant results support
tonly a linear relationship.
I The test of hypothesis Id, regarding media scope and
i
perceptions of threat, showed no support for the
hypothesized relationship. Rather, one could claim
abundant support for the hypothesis that the widest media
scope will be associated with the highest threat
perceptions.
J Hypothesis 2 predicted that when other demographic
jand dependency variables were controlled, the correlations
between education and income and threat perceptions would
be negative and significant. The correlations were
negative, but that of education did not even approach
significance. The correlation between income and threat
i
perceptions was negative and significant. Hypothesis 3,
j
which predicted a significant and positive correlation
between threat perceptions and age, was not supported. In
fact, the relationship between age and threat perceptions
was completely different from city to city-positive in
Austin and negative in San Bernardino—although not
significant in either city.
i
The third and fourth research questions regarding
media use and threat perceptions and dependency relations
provided some interesting and consistent results. Threat
i
perceptions are never significantly related to use of any
of the media examined here, in either city. Certain
^dependency motivations—particularly solitary play and
i
(
social understanding—provide significant predictors of use
for all media.
177
These results will receive closer examination and
more extensive discussion in the following chapter.
178;
i
Chapter Six: Discussion I
j Threat Perceptions and Dependency Relations
This research aimed to examine the proposition that
the perception of a high level of threat in the
environment is linked to more intense media system 5
dependency relations. The perception of threat in the <
environment is seen to be higher among those most j
objectively threatened by American society—those with
lower incomes, women, and non-whites. This observation
provides an indication that threat perceptions are not
based on fantasies about unreal or unlikely hazards but on •
t
relatively "real" possibilities. 1
Three contributions emerge from my research. First,
i ;
■ clear support is provided for Ball-Rokeach's proposition
regarding the relationship between perceptions of threat
i
J in one ' s environment and the intensity and scope of one ' s
I media system dependency relations. Secondly, a
i
j theoretical and an operational definition of perceived
i
(
i threat are introduced which prove useful in this research
and which suggest future research. Finally, an
operationalization of the substantive and media scope of
dependency relations is introduced. This development
(
|
I allows for further support for the ideas that Ball-Rokeach
i
| et al. (1991) and Colman (1990) presented regarding the
nature of media system dependency relations.
The results support Ball-Rokeach's (Ball-Rokeach et ,
al.. 1984; Ball-Rokeach, 1985) proposition that media j
system dependency relations will tend to be more intense
i
I
when threat perceptions are high. Support for this j
proposition has several implications in light of prior
research into dependency relations. It has been noted
i
that the ability of a television program to influence 1
i !
I behavior and belief is enhanced when dependency relations
I
i
| with television are most intense—even when the program is <
j not designed specifically to persuade (Ball-Rokeach et
I aL., 1984; Colman, 1990). My analysis (see Table 5.18)
shows that the positive correlations between threat !
perceptions and television dependency relations for
j understanding and orientation are highly significant.
' Taken all together, it would appear that, due in part to
their heightened perceptions of threat in the environment,
j the socially disadvantaged are somewhat more subject to
i
! the persuasive effects of television than are those who do
i
jnot perceive high levels of threat.
Is this apparent relationship due to a relationship
i
between dependency relations and the structural correlates
of threat? To test this explanation, demographic
( variables were added to the regression analyses used to
j test the relationship between threat and dependency
| relations. In these analyses, threat remained a highly
18o!
significant predictor of the intensity of dependency j
I
relations in every substantive motivation and in every
I
j medium, even when demographic variables were themselves
significantly related to dependency relations. In most j
f
instances, threat was the single most powerful variable in j
jthe equation explaining dependency relations. Ball-
i ;
jRokeach's position regarding a linear association between j
perceptions of threat and intense dependency relations ;
receives strong further support from this analysis.
I Further support for Ball-Rokeach's proposition !
I
appears in the analysis of the substantive scope of
i
j dependency relations and threat perceptions. The trend of ;
i
j the results was linear, although the difference between
j the low and moderate scope groups was not always
i
! statistically significant (e.g. see Table 5.13). In all
I media the highest threat perceptions were associated with
3
' those with the widest substantive scopes. Greater amounts
of perceived threat were associated with a greater
j propensity to evaluate the media as helpful in serving a
! wide variety of personal goals.
j A third indication of support for Ball-Rokeach's
I t
; proposition is the relation between threat perceptions and
1 the media scope of dependency relations. There was rarely
i
| a significant difference between the threat perceptions of
J those with very wide and those with moderate media scopes.
There was virtually always a significantly lower j
\
perception of threat by those with the most narrow media j
scope. Contrary to the findings of Janis (1958), which i
concerned information-seeking at the interpersonal level, j
it appears that those with high perceptions of threat find \
a moderate to wide range of media helpful in the pursuit
I , i
!of their goals.
1
Janis noted that one possible explanation for the
reticence of surgical patients who felt acutely threatened j
>
was their psychological motivation to please their doctors I
I [
land nurses by being "good patients" by not pestering them
for information and reassurance. This motivation, Janis \
suggested, may be the result of unsatisfactory resolution
of the Oedipal complex in childhood, leaving an insecure
I attitude towards authority figures which manifests itself,
I
jin part, as a reluctance to ask for information even when
i ^ ;
it is most desired. Janis suggests that heightened
I perception of threat stemmed from much the same
I
Jpsychological phenomenon, in that neurotics will perceive
i
|themselves to be more vulnerable. Thus the information-
!
jseeking behavior Janis observed and the perceptions of
threat were both attributed to the same root cause. In
the case of mass media, it could be that the fear of
•approaching authority figures is removed or moderated. In
I
I
ithis way, mass media can be "authoritative" without
_ _ (
182;
i
appearing authoritarian, and provide those with acute
I
threat perceptions with sources of information by which ;
i they are not personally intimidated.1
I
The net result of my analyses of threat perceptions j
i
and dependency relations is that a pattern of information ;
seeking from the mass media is discernable which is
I ’
I
, different from that suggested by previous research into i
i human interaction. People's evaluations of the j
I helpfulness of the mass media increase as their
j perceptions of threat increase. This finding is
|
1 particularly consistent for newspapers and television. In
I
j the surgical ward, Janis (1958) noted the tendency of
those experiencing greater threat to ascribe unhelpful
Jqualities to the hospital staff. Boyd's (1989) findings,
!
while not as rich in detailing the personal motivations of
I
' executives as Janis' are for surgical patients, imply that
i
i
\ in the face of a threatening environment those with a low
j 1The need to find unintimidating authority figures in
; order to fight prejudice is put in similar terms by
t Frenkel-Brunswik:
In our present-day struggle to achieve a j
strengthening of the tolerant, liberal point of view ’
we may have to avoid presenting the prejudiced
individual with more ambiguities than he is able to
I absorb. . . . Efforts to modify the "prejudiced"
| pattern may have to make use of authorities—though by
| no means necessarily of authoritarian authorities—in
order to reach the individual in question. This
follows from the fact that it is authority more than
anything else that structures or prestructures the
! world of the prejudiced individual (Frenkel-Brunswik, '
| 1950, p. 486).
tolerance for ambiguity will turn away from environmental
scanning (information-seeking) and focus their attention
on their own organization. Among those in a random
jsample, which may include surgical patients and business
executives, the mass media steadily increase in perceived
helpfulness as threat perceptions rise.
In light of theory and earlier research, some
questions are raised by the relationship between threat
i
perceptions and media system dependency relations. As
I
noted above, perhaps the most important predictor of
^threat perception is the actual probability of a dangerous
I
event (similar to the conception of "risk” offered in
Chapter Two). In times of crisis, when many more people
i
than normal perceive their environment to be threatening,
the mass media may acquire an influential role not
regularly accorded to them (c.f. Hirschburg et al.. 1986).
This is an old assertion, similar to that made by Lippmann
i
(1922) regarding the potential of propaganda in wartime to
produce the "pictures in our heads" which shape our
opinions. The results of the present study, coupled with
the effects process related to media system dependency
I
theory (Ball-Rokeach et al.. 1984; Ball-Rokeach et al..
1990; Colman, 1990), demonstrate that for the subset of
the population for which threat is an everyday perception,
persuasive or propagandistic effects may be more
184;
• commonplace as well. It may be that in a generalized
< j
crisis, the rest of us become more like those who are j
\
already disadvantaged and therefore we adopt their methods !
J
of information seeking (Hirschburg et al.. 1986). j
i
How, then, can we insert threat perceptions into the
model of the media effects process described by Ball- j
Rokeach and tested by Colman (1990)? The results reported I
j in the previous chapter demonstrate that threat
J i
perceptions play are significantly correlated with the !
| i
j intensity and scope of dependency relations in the
I {
I
understanding and orientation motivations, and that threat
perceptions are affected by certain demographic variables |
i
! used by Colman as exogenous variables (i.e. sex,
' ethnicity, education, and income). Present results show
that media use is not a significant predictor of threat,
J or vice versa, so exposure need not be directly related to
threat perceptions in the effects model. I suggest that
perceptions of threat are part of a person's agenda prior
to using the media, and these perceptions contribute to
i
■ heightened arousal to pertinent information both directly
t
j and through the intensifying effect threat perceptions
have on dependency relations. When a study addresses a i
i
| specific concern, as Colman's did for holistic medicine,
, perceptions of threats of a more specific kind may be
included as a variable predicting the intensity and scope
of related dependency relations, and, perhaps, cognitive
and affective arousal to relevant media content.
Ball-Rokeach et al. (1984) and Colman (1990)
demonstrate that "effects studies" may be more fruitful if
dependency relations are controlled in the analysis. The
present results suggest that perceptions of threat in the
environment are important as well, and that those with
higher perceptions and more intense dependency relations
may be the most susceptible of all to whatever media
effects one hypothesizes. The observation that these
people tend to be poorer, minorities, women, and people
with children offers a clear agenda for future research.
I will return to the subject of future research after
discussing two other main contributions of this study and
a more detailed discussion of the other observations
reported in Chapter 5.
A Definition of Threat
The conceptual definition of "threat" I offer in
Chapter 2 represents a contribution to research on this
topic and on related topics. The advances my definition
holds are detailed in Chapter 2, and will be reviewed
here. First, the definition is consistent with earlier
research. The inquiries into stress in the 1950's and
1960's relied on a concept of threat which, while
undefined, substantially corresponds to that which I
offer. The research into "short-circuiting" threat, for
instance, manipulates the dimensions of threat I call
danger and vulnerability in order to show how people
evaluate graphic depictions of physical threats
differently in light of such manipulation. A second
advantage of my definition is that it accentuates the
perceptual nature of threat, distinguishing threats from
iactual hazards. This allows a conceptual distinction from
other concepts, particularly stress and risk, which tend
;to be operationalized in terms of their physical
J
manifestations—e.g. a change in respiration or an increase
in levels of toxic substances. Finally, my definition is
!readily operationalized in terms that most respondents can
junderstand. Asked to evaluate their vulnerability to
jpresumably dangerous phenomena such as muggings and
earthquakes, people seem capable of making reliable
i
jevaluations (see Table 5.1). I believe that further
i
:research can fruitfully explore the dimensions of threat
and the variables which have an impact on perceptions of
i
j threat.
i
j Threat Perceptions and the Scope of Dependency Relations
| This study provides the first operationalization of
I the concepts of media scope and substantive scope of
dependency relations. The operational definitions in
Chapter 4 provide useful and interpretable results which
187
encourage further use of these definitions in dependency
studies. Two strengths of these definitions are (1) their ,
reliance upon relatively strict definitions of "high" and J
{"low"—i.e. relation of the response to the standard
j
deviation, and (2) their derivation from the intensity j
scores, as their theoretical definition suggests (i.e. as [
opposed to constructing a separate scale for the
measurement of scope). The true test, of course, is their i
continued helpfulness in future studies of media system
dependency relations.
The relationship between threat perceptions and the ’
i
desire to acquire information from the media is
1 demonstrated clearly in the analyses of the scope of
| I
(dependency relations. Wider substantive scope and wider !
i ;
media scope were both significantly associated with higher
i
perceptions of threat. For those with particularly high ,
1
j )
|threat perceptions, mass media are considered helpful in
{the pursuit of a wider variety of personal goals than for
those with lower perceptions of threat.
J
| That is not to say that those with lower threat !
perceptions are not searching for information, only that
they are not as likely as their counterparts to perceive
!the media as particularly helpful sources of the
information they desire. Recalling the social and
i
|economic advantages of those with lower threat
188
perceptions, it is not surprising that they may have
i
jgreater access to information resources not under the
control of the mass media presently studied, such as
libraries, professional consultants and therapists, and
even personal acquaintances with more expertise in a
variety of areas (lawyers are more likely to know doctors
personally than teamsters are). It appears that, at least
where mass media are concerned, those with higher
perceptions of threat do not close themselves off to more
!
'sources of information, but quite the reverse.
j
' Another interesting result concerns the substantive
scope variables for radio and newspapers. There are many
jmore radio stations available to a person than there are
!
'newspapers, and once one owns a radio, he or she can
receive radio programming without any additional
expenditure. It would seem that radio is more likely to
i
:provide information helpful to a greater variety of goals
than newspapers. This generality is still subject to the
marketing assessments of sponsors and radio programmers,
but the broader bandwidth of radio provides more room to
address the substantive motivations of more people than
r
'newspapers. This might lead one to expect that the
I substantive scopes of people's dependency relations for
radio would be wider than those for newspapers. Actually,
:crosstabulations of radio and newspaper substantive scopes
show that more often than expected statistically (using
Chi-square, p < .0000) those with low, moderate, or high
j scopes for one medium have identical breadths of scope in
i the other.
As anyone twisting the dial of a radio knows, more
channels does not necessarily mean more diverse
,
j programming. It appears from the present results that
j the capability of one well-financed newspaper to provide
information about a wide range of topics in a given week
is virtually equal to the combined capabilities of a
I city's entire allotment of radio stations. The analysis
described above rests on a comparison between a newspaper
and a radio station. It could be that a more appropriate
comparison is between a newspaper and a radio format. In
i
any given radio market, many stations may share the same
f
i essential format (e.g. talk radio, album-oriented rock
i
; music, etc.). This may limit the actual range of
motivations people may find addressed across the radio
; spectrum.
2This point is demonstrated for television by Waterman
and Grant (1991), who observe that the explosion in
I available channels brought about by cable technology has
i not resulted in a similar explosion in programming
diversity. The economic forces which determine
programming encourage the purchase of syndicated programs
j and movies rather than the production of original
programming.
Threat Perceptions and Media Use
Introducing media system dependency theory in Chapter
1, I noted that simply measuring time spent using the
t media to account for hypothesized effects or other media-
j related outcomes may be minimizing the context of media
behavior. The importance of a more complicated concept of
media relations is demonstrated by the regression
j equations where use is the dependent variable. If one
| were to expect that threat perceptions would predict more
use rather than more intense dependency relations, one
would find no support for that hypothesis. Threat is not
significantly related to time spent using any of the four
; media studied when dependency relations and demographic
j variables are controlled. In all instances save newspaper
j readership, the single best predictor of use was the
i
; intensity of the respondent's solitary play dependency for
that medium.3 This is not surprising or contrary to media
j system dependency theory (which offers no particular
i
j expectations regarding frequency of use). The
j insignificant relationship between frequency of television
! viewing and threat perceptions does challenge cultivation
l
j theory, however.
3For newspapers, the intensity of the social
I understanding dependency relation was best, followed
I closely by the solitary play relation (see Table 5.26).
If the "enculturation" hypothesis were to be j
supported in light of these findings, then it would be I
J
! necessary to demonstrate that those with the highest
i
i
threat perceptions, (who happen to be socially
disadvantaged in one way or another), would be the most |
avid consumers of television. There is little support for i
I
i |
this hypothesis. When the variables which predict
television viewing are examined, one finds that the two !
most significant demographic influences are being
j
: unemployed and having a lower income. Threat perceptions, ,
| however, are not significantly related to television i
j viewing (in fact, the regression coefficient is negative).
When the relationship between television viewing and
threat perceptions is examined from the other direction,
i.e. using use of television to predict threat
: perceptions, the results are similar. Controlling for
I
I demographics, dependency relations, and use of other
I
1 media, I found no significant relationship between
; perceptions of threat and television viewing—or between
I perceptions of threat and use of anv medium. The
significant variables are the familiar demographics—sex,
! ethnicity, children, and income—as well as the intensity
of the action orientation dependency relation for
i television. If threat perceptions do not predict use and
j use does not predict perceptions of threat, there is very
little indication that use of television leads one to
jbelieve that he or she lives in a "scary world" (Gerbner
and Gross, 1976). The present findings are consistent
with the reanalysis of cultivation studies offered by
Hirsch (1980), with the attempted replication of Gerbner
et al.1s studies by Wober (1978) (who found no cultivation
effects in Britain when similar statistical controls were
introduced), and with the findings of Doob and MacDonald
|(1979).
I
This is not to say that the distortions of reality
I
|that Gerbner et al. note through their content analyses
!are not important or do not contribute to perceptions of
i
the world. What is important to observe is that it is not
J necessary to be a heavy viewer to be affected by
j television content, but to be viewing with a substantive
motivation that heightens arousal to the material (Ball-
I
Rokeach, 1988; Colman, 1990). It has been demonstrated
above that high perceptions of threat are associated with
more intense dependency relations in nearly all
|substantive motivations. To the extent that those who
I
'feel themselves to be threatened day to day also develop
i
(intense dependency relations with television in order to
I gain information which they believe will help them deal
!
'with the threats they perceive, then any distortion of
content may have a measurable effect not discernable in
the present study—and certainly not discernable in studies j
!
which do not account for the context of dependency
relations.
The relation between threat perceptions and
l
1
dependency relations proposed by Ball-Rokeach is based on ‘
1
the premise that most people, in the face of what they
I
perceive to be threatening circumstances, will seek to j
gather information in order to define the situation or to !
learn what to do in the face of the threat. The acute
I need for information makes the resources controlled by the :
|
media that much more valuable, and therefore the intensity
of the relationship increases. The results of the present
study, particularly for newspapers and television, support!
< this position.
I Dependency Relations with Newspapers
^ Twenty years of research into the readers of
; newspapers was summarized by Bogart: "[Subscribers] are
J people of above-average income, education, and age; people
who are married, homeowners, and established in their
communities" (Bogart, 1981, p. 46). While Bogart refers
to subscribers, the profile he offers is echoed for
readers in general by Burgoon and Burgoon (1980). Present
I
. results demonstrate that, controlling for dependency
i ■
| relations, higher income is not a significant predictor of
I
J newspaper use. Higher income, controlling for threat
perceptions, ethnicity, and sex, is a significant
predictor of higher social understanding dependency
relations, which in turn is a significant predictor of
use. A decline in newspaper reading in the last two
decades may not be directly related to stagnant real
income in the middle class, or to the flight of
traditional readers out of the central cities, but to
changing perceptions of the helpfulness of the newspaper
to meet goals of social understanding.
Two substantive motivations, social understanding and
solitary play, provide the strongest predictors of
newspaper use. Since solitary play dependency predicts
use for every medium studied (and since intuitively the
helpfulness of mass media for goals of individual
amusement seems quite likely to explain a great deal of
use), in the case of newspapers it is most interesting to
focus on the variables related to the social understanding
dependency relation in order to gain insight into reading.
Loges and Ball-Rokeach (1991) hypothesized that the
intensity of the social understanding dependency relation
would predict newspaper use. The theoretical basis for
this hypothesis was that the socially integrated person
Bogart (1981) described (see the quotation above) seemed
most likely to require the sort of complex and constant
update about world and community affairs that the
195
editorial content of newspapers provides. In other words, !
affluence would lead to a more intense goal of social
i
understanding, which in turn would lead to greater use. j
j I
! This hypothesis was supported (Loges and Ball-Rokeach, |
i
1991). The present results demonstrate that another
reason for an intensified need for social understanding is !
perception of threat, and that those with high perceptions ,
j of threat tend not to be those with the material concerns |
\ |
of the affluent. I
It is here that a distinction between use and
dependency relations for newspapers can be drawn. The
affluent have been shown to use newspapers more than
others in previous research which did not control for
i
dependency relations. The less affluent can be shown to
l have the same dependency relation (social understanding)
I
i
| with newspapers as the affluent, albeit for different
i
I reasons, but not the same patterns of use. Both groups
i
recognize the value of the medium for addressing their
goals, but their actual habits of use are demonstrably
I
j different.
j Dependency Relations with Television
| The results of the analyses for television resembled
those for newspapers. Solitary play and social
understanding motivations, respectively, offer the best
predictors of use. As threat perceptions rise dependency
|
i
r
relations of understanding and orientation tend to be more ,
i
intense. '
Television use, more so than use of any of the other j
i
media, is significantly associated with lower social |
status even after threat perceptions, dependency j
relations, and use of other media are controlled. The i
i
unemployed (whether looking for work or not), the less j
educated, and those with lower incomes report
significantly more television viewing than others. Of |
i \
| course, it could be that those with higher status are more
reluctant to claim high amounts of viewing, and this must
be considered a plausible explanation for some of the \
variance between the classes. Robinson (1981) notes as
well that "the main factor drawing people away from i
i ;
j television is the attraction of activities outside the
I home" (Robinson, 1981, p. 127). If the area one lives in
1
| discourages outside activity, or if outdoor amusements are
! not easily afforded, one may be more likely to seek
|
amusement at home.4
j As was the case with newspapers, despite the
significantly more intense dependency relations associated
i
i
with high threat perceptions, and the significant positive :
j 4Robinson also notes that the fact that television
viewing and newspaper use are both "home" activities
probably accounts for much of the correlation between the
! two.
| 197
i
; relation between social understanding dependency relations
!
I and use, threat perceptions themselves are not
i
significantly related to use. In fact, while
insignificant statistically, the relationship between
i
threat perceptions and use of newspapers and television is
negative when the significant variables are controlled.
Some Anomalous Findings for Radio and Magazines
l
!Radio
I
j In the combined sample, the regression equations for
I
I radio provide strong support for the proposition that the
i
i
|intensity of dependency relations increases as threat
'perceptions increase. In no substantive motivation is the
; non-linear equation a significant improvement on the
:linear equation. This overall finding, in combination
with those for the other media, lend support to the
generalized relationship between the social environment
I and media system dependency relations proposed by Ball-
1
|Rokeach (1985).
I
Certain anomalous findings in the city samples
I demonstrate the impact of other social forces on
I
dependency relations. The linear relationship between
i
!
|threat perceptions and media system dependency relations
!which is observed generally is not observed for the city
iof Austin in the self understanding and interaction
;orientation substantive motivations. Instead, the
198
. i
[hypothesized parabolic relationship is observed.3 It is ;
noteworthy that this relationship is fundamentally t
i i
different from that observed in most other instances in
the present analyses.
I
Are these anomalies the result of chance? There were [
I
| 32 separate tests of significance in the regression
i :
J equations for each sample (4 media x 4 substantive
|motivations x 2 equations). Since these findings were
i
I significant only at the .05 level, there is a case to be
| ]
I
| made for a finding due only to probability (in 40 such
[tests, 2 errors of this sort would be expected at the .05
ilevel).6 While it is tempting to explain anomalies as the
j inevitable result of tossing the statistical coin so many
;times, it seems fruitful to attempt to explain this result
i
ion its face as well.
Why would the people who perceive higher amounts of
i
| threat in Austin regard radio as less helpful for goals of
i
j self understanding and interaction orientation than those
| 5It will be recalled from the factor analysis of
j dependency scale items reported in Chapter 5 that items
!designed to measure the self understanding and interaction
!orientation motivations intercorrelate relatively highly,
jIt is therefore not surprising that a relationship
iobserved for one is seen for the other as well.
6It should also be noted that the sample size,
; particularly in the combined sample, is large enough to
iproduce statistical significance with fairly low beta
icoefficients. In the separate city samples, this problem
i is reduced of course.
199]
who perceive only moderate amounts of threat? One ]
possible explanation is supported by the greater affluence i
of the Austin sample than the San Bernardino sample and by !
I
the status of Austin as a state capital and education
center. Each of these differences between cities may
i
account for differences in the media agendas between the
jcities.
i
I Ball-Rokeach (Ball-Rokeach et al.. 1985) proposed
l F
f
that the media agenda is one social antecedent of media f
system dependency relations. The sorts of social crises
examined as threats in this study are often expected to be ,
I
dealt with by government and science. In a city such as
i
:Austin, where government officials and research scientists ;
^ reside, treatment of these issues by the media may be
! different than the treatment afforded those subjects in a
I
! less "capital'1 city such as San Bernardino. To the less
I
:affluent and less powerful radio audiences in Austin—who
perceive these social problems as higher threats—the
J
j public policy or social science perspectives may catalyze
social understanding dependency relations (which are
j correlated in the linear form), but not offer the most
I
1 helpful avenue towards self understanding.
, In San Bernardino, a U-shaped parabola fits the data I
for the social understanding dependency relation and
threat perceptions, such that those with low and high
200
threat perceptions develop more intense dependency
relations while those with moderate threat perceptions
I
I
j tend towards dependency relations of lower intensity.
I
j This observation is, of course, directly contrary to the
i
j hypothesized relationship, and not predicted by the
alternative proposition of Ball-Rokeach of a linear
relationship. This finding implies that the social
i
| understanding goals of those with low and high threat
J
iperceptions have something in common that radio
1
programming in San Bernardino addresses.7 Let us suppose j
I that, in a city where politics is not an unusually central
i
concern (as it is in the Texas state capital, Austin),
public affairs programming on radio is less likely to be
immediately relevant. Perhaps a more detached
; presentation of the news and other issues-oriented
j information appeals to those with low threat perceptions,
j and allows those with high threat perceptions to accord
more credibility to radio reporting. Those with moderate
;perceptions of threat—i.e. most people—might consider such
i '
j a presentation to be trivial and not a very helpful
’ treatment of social affairs.
i
I
J
I
i
I ----------------------
i
7N.B. This is not to say that these two groups listen
to the very same programs on their radios, only that the
j medium in general is evaluated as helpful by those with
i low and those with high perceptions of threat.
The results of the present study and the explanation
offered suggest further study of the differences in the
media systems of both cities, and of the relationship
between the media agenda and dependency relations.
Further study would no doubt seek to account for the
differences in driving habits which may exist between San
Bernardino—which is part of Southern California's "car
culture"—and Austin, since radio listening is an activity
associated with driving.
Magazines
Of the four media studied in this project, magazines
clearly offer the best opportunity for one to regularly
obtain information about a narrow topic. Publications
devoted exclusively to skateboarding, conservative
politics, Christian Science, soap opera news, or
psychology appear weekly or monthly and can be readily
found on supermarket magazine racks. This demonstrable
diversity actually ought to make prediction about media
behavior easier, since people can more readily tailor
their magazines to their other personal attributes.
In light of the above comment, it is not surprising
that the only support seen for hypothesis la in the
combined sample can be found in the analysis of magazines,
where the self understanding motivation is related to
threat perceptions in the hypothesized parabolic fashion.8
!
j The Austin sample provides similar support for Hla for
self understanding, but in San Bernardino the test of the
hypothesis only approaches statistical significance (p =
.0646).
The ability of a person to create a "personalized"
' information environment by reading a specific magazine or
!
iassortment of magazines provides magazines, more so than
I
I other media, with an opportunity to assist self
understanding goals. One's choice of a public affairs
jmagazine can simultaneously address one's social
! understanding and self understanding goals—e.g. reading
I
The Nation as opposed to the National Review. To observe,
then, that the trend towards linear relationships between
i
I threat perceptions and dependency relations does not apply
1
to the self understanding motivation for magazines might
|suggest that the closer one gets to studying people's
preferences for content in addition to media forms, the
! more one may see the presently hypothesized patterns of
media behavior.
Perhaps, however, magazines are appraised to be less
I helpful for self understanding by those with high threat
i
! 8As with the anomalous results for radio discussed
j above, it must be noted that chance alone may account for
!this result. In this case, the significance level was
I.01, but this result is obtained in the combined sample,
;with a sample size of 935.
203
perceptions because they cannot generally afford to ,
include magazines in their media diet. Lack of exposure I
i 1
{ to a medium could certainly lead to an appraisal of the
i '
I medium as "unhelpful” for meeting one's goals. The j
I
i
regression equation explaining magazine use does show a
negative regression coefficient for self understanding,
I
jbut it is not significant. The coefficient for income was
i
j positive, but not significant. Lack of money to buy
j magazines may play some small part in the evaluation of
l
magazines offered by those with high threat perceptions, j
j but income does not appear to be a significant obstacle to
i reading magazines, particularly when other demographic
!
factors not related to threat perceptions are controlled.
j
’ One other significant non-linear solution appeared in
j the regression analyses for magazines. In San Bernardino,
I
(
; the linear term for threat was not significant, but the
j
squared term was (see Table 5.16). This equation does not
: define a parabola, but rather the values for social
| understanding dependency continue to rise geometrically
and then taper off somewhat—as they must, since the
; ceiling forced on them by the boundary of the scale is
I
|reached. In other words, among the San Bernardino
{ respondents, higher threat perceptions lead to very high-
,intensity social understanding dependency relations with
■ magazines. Why? Perhaps the differences between the San
204
I Bernardino and Austin samples could provide some insight
into this relationship.
It will be again recalled that the respondents from
San Bernardino were less affluent, less well educated, and
I
i
| disproportionately female. In short, the respondents from
; San Bernardino fall into the profile of those with high
j threat perceptions more so than those in Austin (thus
i
I
accounting somewhat for the higher mean perceptions in San
Bernardino—i.e. that difference is due at least in part to
sampling error). The unusual relationship between threat
perceptions and social understanding dependency for
;magazines in San Bernardino may not have much to do with
j
1 the city of San Bernardino per se. but with the
I
demographic correlates of high threat perceptions.
That said, what about the demographic correlates of
ihigh threat perceptions might make magazines so helpful in
I
'addressing social understanding goals? Income and
■ ethnicity are not significantly related to readership in
San Bernardino (or in Austin, once use of other media is
j controlled), but sex is (men read more). Perhaps
J magazines are too infrequently directed to the self
( understanding goals of poor minorities, and so the
j magazine content they do read is likely to be seen as
!pertinent information about the rest of society—i.e.
i
,contribute to social understanding goals rather than to
205;
I
self understanding goals. If that were the case, we would j
expect exactly the relationship we observe—that those to !
I
whom magazines tend not to be personally relevant consider
magazines more helpful for social than for self j
i
understanding. Given higher threat perceptions and |
; roughly equal time spent reading, the logic of the I
j |
i explanation offered for the S-shaped curve for social 1
! !
understanding in San Bernardino provides the explanation
for the bell-shaped curve for self-understanding.
The overall pattern of results for these four media \
i
provides some interesting insight into the differences j
i
between them. The two media'with the most narrow range of
I
alternative "channels"—newspapers and television—are the
i two which most consistently support Ball-Rokeach1s
i
: proposition regarding the impact of threat perceptions on
I
J dependency relations. Radio and magazines, however, which .
! offer the widest range of distinct frequencies and
[
publications, are not perceived as helpful for the pursuit
of self understanding goals by those with higher threat
\ ,
I
, perceptions. It may be that the economics of radio and
I 1
j magazines, which make possible the relatively large number
I |
! of separate, more or less specialized outlets (more for
I
| magazines, less for radio), also result in the exclusion
from the "information market" those who are disadvantaged
i
]and threatened by their social position. A magazine
206|
targeted at a particular interest group or demographic j
!
group is working with a far narrower margin of error than
j a television station targeted at virtually anyone with a
! television set. As Waterman and Grant (1991) have pointed !
I |
iout, even the opening of the cable spectrum to more
I
| channels does not seem to have significantly narrowed the
!
j audiences sought by programmers. Radio and magazines may ^
I be perceived as less helpful to the disadvantaged with j
j high threat perceptions because commercial radio stations
j and magazines cannot afford to be helpful to these people.
i
I Weaknesses in the Present Study
i
| The relatively low response rate in San Bernardino—
whatever accounts for it—encourages caution in
i interpreting results peculiar to that city. These include
several anomalous significant non-linear equations (see
i
;Tables 5.14 and 5.16, and the discussion above). On the
whole, however, the results between the cities are
similar, particularly the trend of the substantive and
| media scope analyses. Also, the disproportion provided
■ some face support to the correlation between threat
perceptions and lower socioeconomic status, since threat
i perceptions appear higher in the San Bernardino sample.
| Another weakness, however, is implied by the
i
; "appearance” of higher threat scores in California—the
( response scale was not the same in each city. While the
response rate may be attributed to any number of causes,
some out of the researchers' control, the difference
between the scales is due to a lapse in coordination
between the sites. For analytical purposes, this has not
proven to be a significant problem.
The analyses in this report are compromised somewhat
by the necessity to correct the distribution of the
dependency responses in order to apply a regression
analysis to these variables. In the newspaper and
television analyses, the effects of this compromise are
not as great as those in the radio and magazine analysis.
This is because the newspaper and television analyses
showed no departure from the tendency of threat
perceptions to rise with the intensity of dependency
relations in any substantive motivation.
In radio, however, where the Austin self
understanding and interaction orientation motivations
showed significant values for the coefficients for the
squared values of threat, the self understanding and
interaction orientation variables were composed of 15% of
scores for the newspaper action orientation variable.
Given the pattern of relationships observed, this may have
muted the true significance of the regression coefficients
for the non-linear terms somewhat. In magazines, the
interaction orientation variable similarly consisted of
208
15% of the newspaper action orientation variable. Since
the self understanding and interaction orientation
dependency scales were so highly intercorrelated for
magazines (and all other media), this compromise may
i
explain why, in the combined sample and in Austin's
sample, the interaction orientation variable's non-linear
regression coefficient was not significant even though, in
i
the same samples, the term for self understanding was
significant. In other words, the anomalous finding was
!
that of an unweighted variable—magazine self
|
!understanding. It is possible that the weighting of
magazine interaction orientation with a variable that
!showed no significant non-linear relation to threat
'suppressed any significant relationship—i.e. a further
I
!anomaly—which otherwise may have been observed.
I
On several occasions, I have remarked that the
present data do not allow me to calculate the "true"
!probability of the various threatening phenomena specified
|in the scale coming to pass for the respondents. This
I knowledge would have added an extremely helpful control
i
variable to those already gathered, and allowed some more
lextensive exploration of the concept of threat perception.
Crime statistics (c.f. Doob and MacDonald, 1979),
epidemiological studies, geological surveys (c.f. Turner
i
i
|et al.. 1986), and economic indicators can allow some
209;
accuracy in the calculation of the vulnerability of a
population, but in the end the number of individual ,
I
I
deviations from the norm in these matters, especially over '
such a wide range of concerns, make it daunting to account '
for a reasonable amount of the variance in vulnerability.
For each respondent, his or her exposure to each threat
!
would have to be calculable. This would involve a j
!
j
considerably longer and more intrusive survey instrument
(particularly for health-related items). Nonetheless, ■
|
j future research might benefit from concentration on one or :
(another of these threats, thereby benefitting from
!
!knowledge about the risk of exposure to specific threats.
!
I
Other suggestions regarding future studies will now be
considered.
! Suggestions for Further Research
; The association of higher perceptions of threat with
j those who are most likely to suffer from the social and
j economic structure of the United States suggests several
possible avenues for further study. The interpersonal
i
information behavior of those with higher perceptions of
threat is a central area to explore with the present
findings in mind. What sort of dependency relations exist
at that level, beside the mass media system? Who is
t
considered helpful? What community groups provide
:information which competes with the mass media for the
attention of those with high, moderate, or low threat j
perceptions? The mass media system exists alongside other
sources of information, and the "effects" of anv of these
i
sources are best understood by accounting for the others. J
Ball-Rokeach (Ball-rokeach et al.. 1984; Ball-Rokeach
j
et al.. 1990) has proposed that, following exposure to a
mass media message, the interaction one has with other |
information sources can play a significant role in the |
processing of the mediated message. The interaction
between interpersonal and mass media dependency relations
has been explored by Aydin et al. (1991). Colman (1990)
included interpersonal communication about holistic health
in her study. Dependency relations with people and
institutions which control information resources outside
the media system—such as one’s coworkers, one's doctor,
one's neighbors, one's boss, or the local police force—can
combine with media system dependency relations to produce
one's interpretation of the message and guide one's action
i
following reception of the message. Research that seeks
to "map" the web of dependency relations in a group or
community—both media and interpersonal relations—could
enhance our understanding of the effects and nature of
these relations.
The discussion above of the perceived unhelpfulness
of radio and magazines to address the self understanding
211
goals of those who have higher perceptions of threat
I
provides more research questions. How are the content of
specific magazines and radio programs understood or
| interpreted differently by people with different j
!
1 perceptions of threat, or of different socioeconomic
i
status? What are the program-level or publication-level
information habits of different groups, i.e. do they read I
i
different magazines, listen to different stations, watch 1
different television programs, and read different
newspapers? How much of the difference in dependency
relations for a media form can be attributed to these \
^ differences at the level of content consumption? 1
Beside the variables studied presently, what
icontributes to perceptions of threat in the environment?
i
; Are more educated people better able to determine the
i
I actual potential for harm (i.e. is their conjectural ;
f
component of threat less subject to error)? How may risk, ,
as it is defined in Chapter 2 for present purposes, be
taken into account when people's threat perceptions are
j interpreted? It would be interesting to know the
information habits of those who perceive much more threat
than there is "risk” in their environments, and compare
their habits to those habits of people whose threat scale
;scores are similar, but whose deviation from their actual
exposure to harm is not as far. This psychological
j dynamic is not examined in the present study. It surely
!
|deserves attention, particularly in light of Janis'
contribution to our understanding of the different
information goals and motivations that people have.9
The implications of previous studies of persuasion
and media system dependency relations (Ball-Rokeach et
j al.. 1984; Colman, 1990), coupled with the present
I
findings, provide further direction for research. Given
,the more intense dependency relations of those with higher
jthreat perceptions across media forms and substantive
jmotivations,10 are these people more or less subject to
! persuasive effects than others? Their desire for
information may suspend or enhance the critical faculties
!they might normally bring to bear on information they
[receive. If one fears he or she may lose a job,
information which supports or which contradicts that
suspicion may be clung to with more-than-usual conviction—
i
;or examined with more-than-usual caution. The heightened
icognitive and affective arousal to pertinent information
I
{noted by Ball-Rokeach et al. (1984) and Colman (1990) when
dependency relations are activated has been shown to
I
i
| 9Janis was able to control for this somewhat by
icomparing people who were awaiting operations of roughly
j the same danger, and whose physical conditions were
|similar.
I 10With the notable exceptions for self understanding
!for radio and magazines already discussed.
enhance the possibility of changes in attitudes and 1
behavior. The impact of threat perceptions on such
arousal, and on beliefs and behaviors related to arousal,
should be explored further.
i
| The concept of threat perceptions can be added to the j
i I
imodel of media effects suggested by media system
i |
| dependency theory, and further tests of that model may be
j more fruitful as a result. The present results offer an
I
1 avenue between social location and dependency relations
which runs through perceptions of threat. Ball-Rokeach
has proposed that ambiguity is another concept which may |
l
; play a similar role. Further research into the interplay
'between threat, ambiguity, and dependency relations may
| draw more finely the path from social location to media
j effects as it runs through media system dependency
i
|relations.
i
| There is much to be gained by extending the analysis
{ of threat perceptions and dependency relations over time.
I
j The social forces which affect both are very dynamic, and
certainly causally related to some extent (e.g., consider
the crime rate and the media agenda). An important
expansion of the present analysis would be to assess the
t
I
J changes in social and solitary play dependency relations
! i
as threat perceptions change. Ball-Rokeach (1973) noted
! that the subjects in her study of pervasive ambiguity
214|
/
would engage in information-seeking behavior as a problem- !
'
!solving endeavor, then back away for a while and seek to !
i
;reduce the tension brought about (in part) by the strange
! i
| situation, then return to the task of resolving the j
ambiguity. A similar pattern may be observable over time !
I
! in dependency relations. Those with high threat
;perceptions may experience a cycle of changes in their !
j I
! personal goals from understanding and orientation to play, j
i
then back again if the perception of threat persists.11
j In general, the information-related goals of those
|who perceive themselves to be threatened by their social,
j economic, or natural environments may lead those people to i
;receive, and subsequently interpret, information in a way
I that is different from the way others treat the same
information. This is a similar position to that of Hall
(1981) and Morley (1980), who attempted to theoretically
I
. and empirically account for the effects of ideology on
!
,people's understanding of mass media messages. People in
i
different class locations may have different material
j needs for information, but the mass media are not
I
j structured to address that diversity of needs per se. The
I would emphasize that the pattern of information-
seeking/backing off/information-seeking suggested here
does not imply any dramatic fluctuation in time spent
using the media. The temporal use variable could remain
1 constant as the motivations for use undergo considerable
I change.
215
separate classes must each contend with the information
the mass media do provide—contend with it either by
supplementing it with information from elsewhere, or by
i
grappling with it as it is and making sense of its
I |
I contradictions and inadequacies as best they can. The ;
; !
more one's personal informational resources can make up ,
i I
for the limitations of the media system, the better
jprepared one is to live with the mass media's
ishortcomings. The more intense one’s dependency relations
i
jwith the mass media, however, the less one can afford to
|believe that the mass media are not actually helping them.
Those with high threat perceptions and intense media
system dependency relations may be reluctant to look the
|gift horse of the mass media too closely in the mouth. As
I
;a result, they may be more readily persuaded that what
comes from the horse's mouth is credible.
Conclusion
This analysis began with the proposition that threat
'perceptions may not be related to mass media behavior in
| as straightforward a manner as certain theories suggested. ,
!Overall, the results of this investigation do not support
I
jthat proposition, but the results do present a somewhat
imore complicated relationship between threat and the media
than has been previously articulated. Part of that
complication is introduced by the social characteristics
of those perceiving threat, and further complication is
introduced by the knowledge we have of the consequences
associated with more intense media system dependency
I relations.
' Several contributions to existing knowledge of the
nature and the measurement of media system dependency
|relations are made in this study. A basic proposition
| regarding the role of threat perceptions in shaping
{dependency relations was strongly supported, not only for
i
i
the intensity of relations, but their scope as well. The
*
;measures of scope offered here represent the first attempt
to isolate and measure these variables. The results are
^ encouraging, and further development of these concepts can
t
j continue with at least one method of operationalization
I
|available.
I believe that the strong association between the
concept of threat I presented in Chapter 2 and the people
! whom one would intuitively expect to experience threat in
!
contemporary American society provides reason for
I confidence in the validity of the concept and of the scale
used to measure it. The concept of threat has played an
important role in mass media theory, as well as in
psychological theory. A reliable scale for measurement of
|threat perceptions represents a contribution of this
research project.
217
Stuart Hall has suggested that the level of awareness
j
one has of one's class position—i.e. one's class ;
, consciousness—is likely to have a powerful impact on one's ’
i
t
j acceptance and interpretation ("decoding") of the content
of the mass media. In the present analysis, I believe |
that one's consciousness of oneself as threatened by ^
identifiable social hazards—such as the cost of medical j
care or an unstable economy—when it is not accompanied by !
i
| some awareness of the structural origins of these threats
I
| as they relate to one through class, may have an equally
:powerful impact on one's acceptance and interpretation of
media information. My fear is, however, that perception
jof threat without class consciousness—and the
■ consciousness of the media system's role in the economy
I
I which class consciousness may provide—may only intensify
one's dependency upon the very structure which threatens
one's well-being.
The disadvantaged apparently perceive the danger they
|face; and they know that the next step is to seek
| information about the causes of the danger and the action
, they may take to avert its consequences. Ball-Rokeach
| correctly asserted that this combination of perceptions
I
would be found among those with intensified relations of
dependency upon the society's informational resources.
i
I
i The power of those who control those resources to maintain
218
the social structure which in turn produces the danger is .
truly an awesome display of the mass media's potential to
i perpetuate ideology, behavior, and belief. 1
219
REFERENCES
Aydin, C.E., Ball-Rokeach, S.J., and Reardon, K.K. (1991).
Mass media resources for social comparison among
breast: cancer patients. Paper presented to the
! Annual Convention of the International Communication
I Association, Chicago, IL.
i
; Ball-Rokeach, S.J. (1973). From pervasive ambiguity to a
definition of the situation. Sociometrv 36:378-389.
I ----- (1985). The origins of individual media system
| dependency: A sociological framework. Communication
i Research 12:485-510.
i
----- (1988). Media system dependency theory.
Unpublished manuscript. University of Southern
California.
-----, Rokeach, M., and Grube, J. (1984). The Great
American Values Test: Influencing Behavior and
Belief Through Television. NY: The Free Press.
I ----- and Rokeach, M. (1987). Contribution to the future
study of public opinion: A symposium. Public
Opinion Quarterly 51:184-185.
|
-----, Power, G.J, Guthrie, K.K., and Waring, H.R. (1990).
Value-framing abortion in the United States: An
application of media system dependency theory.
' International Journal of Public Opinion 2:249-273.
; Bandura, A. (1977). Social Learning Theory. Englewood
! Cliffs, NJ: Prentice-Hall.
i
Basowitz, H., Persky, H., Korchin, S., and Grinker, R.
(1955). Anxiety and Stress. NY: McGraw-Hill.
i
Beniger, J.R. (1986). The Control Revolution. Cambridge,
MA: Harvard University Press.
: Blascovich, J., Veach, T.L., and Ginsburg, G.P. (1973).
Blackjack and the risky shift. Sociometry 36(1):42-
; 55.
Blumer, H. (1962). Society as symbolic interaction, in
A.M. Rose (ed.), Human Behavior and Social Processes.
Boston: Houghton Mifflin.
220,
Bogart, L. (1981). Press and Public. Hillsdale, NJ: L. j
j Erlbaum and Associates, Inc.
; Boyd, B. (1989). Perceived uncertainty and environmental
{ scanning: A structural model. Unpublished doctoral
j dissertation, University of Southern California.
! |
l Burton, I., Kates, R., and White, G. (1978). The
Environment as Hazard. London: Oxford University ;
’ Press. !
; Cancian, F. (1979). A useful distinction between risk and i
j uncertainty. Proceedings of the Workshop on Socio-
i Economic Constraints to Development of Semi-Arid
Agriculture, ICRISAT, February.
| Colman, W. (199 0). Health moves to prime time:
i Evaluating the impact of a prime-time television
j movie of the week on viewers' content relevant health
j beliefs. Doctoral Dissertation, University of
| California, Los Angeles.
i
j Cooley, C.H. (1902). Human Nature and the Social Order,
j NY: Scribner's.
I Daft, R., Sormunen, J., and Parks, D. (1988). Chief
| executive scanning, environmental characteristics,
j and company performance: An empirical study.
I Strategic Management Journal 9:123-139.
i
DeFleur, M.L. and Ball-Rokeach, S.J. (1989). Theories of
Mass Communication (5th ed.). NY: Longman.
Dillman, D. (1978). Mail and Telephone Surveys: The
Total Design Method. NY: Wiley and Sons.
I
' Doob, A.N. and MacDonald, G.E. (1979). Television viewing
| and fear of victimization: Is the relationship
i causal? Journal of Personality and Social Psychology
37:170-179.
j Douglas, M.T. (1985). Risk Acceptability According to the
j Social Sciences. NY: Russell Sage Foundation.
----- and Wildavsky, A. (1982). Risk and Culture.
! Berkeley, CA: University of California Press.
I
1 Emerson, R.M. (1962). Power-dependence relations.
American Sociological Review 27:31-41.
221'
Evans, S.H. and Clarke, P. (1983). When cancer patients 1
fail to get well: Flaws in health communication, in j
R.N. Bostrom (ed.), Communication Yearbook 7. |
Beverly Hills, CA: Sage. j
Freimuth, V., Hammond, Edgar, and Monahan, J. (1990). !
Reaching those at risk: A content-analytic study of
AIDS PSAs. Communication Research 17:775-791.
Frenkel-Brunswik, E. (1950). Dynamic and cognitive !
personality organization as seen through interviews, |
in T.W. Adorno, E. Frenkel-Brunswik, D.J. Levinson,
and R.N. Sanford (eds.), The Authoritarian j
Personality. NY: Harper and Bros. i
Garfinkel, (1963). Conception of and experiments with
"trust" as a condition of stable concerted actions,
in O.J. Harvey (ed.), Motivation and Social
Interaction. NY: The Ronald Press Company.
! Gerbner, G. and Gross, L. (1976). Living with television: j
I The Violence Profile. Journal of Communication 1
! 26(2):173-199.
i
: -----, Gross, L., Morgan, M., and Signorielli, N. (1980).
| The "mainstreaming" of America: violence Profile No.
| 11. Journal of Communication 30(3):10-29.
I
| Grant, A.E., Guthrie, K.K., and Ball-Rokeach, S.J. (1989). ,
i Media system dependency theory, television shopping, !
and television shoppers: Theoretical significance
and empirical test. Paper presented at the Annual
Meeting of the International Communication
; Association, New Orleans, LA.
I
I Guthrie, K.K. (1989). Mass media and the competition to
| control value-frames. Predissertation paper,
, Annenberg School for Communication, Los Angeles, CA. ;
I i
i
Hall S. (1981). Encoding/decoding, in S. Hall, D. Hobson,
A. Lowe, and P. Willis (eds.), Culture. Media.
Language. London: Hutchinson. !
i
j Hirsch, P.M. (1980). The "scary world" of the nonviewer
! and other anomalies: A reanalysis of Gerbner et
| al.*s findings on cultivation analysis,
i Communication Research 7:403-456.
Hirschburg, P.L., Diliman, D.A., and Ball-Rokeach, S.J. j
(1986). Media system dependency theory: Responses
to the eruption of Mount St. Helens, in Ball-Rokeach, \
S.J. and M.G. Cantor (eds.) Media. Audience. and i
Social Structure. Newbury Park, CA: Sage. !
i
Horvath, A.T. (1990). MSD survey confirmatory factor ;
analysis: Preliminary results. Unpublished paper,
Annenberg School for Communication, Univerity of
Southern California.
Janis, I.L. (1958). Psychological Stress: Psychoanalytic
and Behavioral Studies of Surgical Patients. NY:
Wiley.
i
Katz, E., Gurevitch, M., and Haas, H. (1973). On the use
of the mass media for important things. American \
Sociological Review 38:164-181.
-----, Blumler, J.G., and Gurevitch, M. (1974).
Utilization of mass communication by the individual,
in Blumler, J.G. and E. Katz (eds.) The Uses of Mass j
Communication: Current Perspectives on
Gratifications Research. Beverly Hills, CA: Sage.
Klaidman, S. (1991). "Muddling Through" in The Wilson
Quarterly. Spring, 1991.
Knight, F. (1971). Risk. Uncertainty, and Profit.
Chicago: University of Chicago Press.
Lain, L.B. (1986). Steps toward a comprehensive model of
newspaper readership. Journalism Quarterly 63:69-74.
Landau, S.F. (1989). The effect of objective social
stress factors on subjective perception of well-being
and social solidarity—the Israeli case. Human
Relations 42 (6):487-508.
Lazarus, R.S., and Alfert, E. (1964). Short circuiting of
threat by experimentally altering cognitive
appraisal. Journal of Abnormal and Social Psychology
69:195-205.
----- , Opton, E.M., Nomikos, M.S., and Rankin, N.O.
(1965). The principle of short-circuiting of threat:
Further evidence. Journal of Personality 33:622-635.
Lippmann, W. (1922). Public Opinion. NY: MacMillan.
Loges, W.E. (1990). The effect of dependency relations on
newspaper readership: Theoretical considerations and
empirical test. Paper presented to the Annual
Meeting of the American Sociological Association,
Washington, D.C.
----- and Ball-Rokeach, S.J. (1991). Dependency relations
and newspaper readership. Manuscript under
submission, Annenberg School for Communication,
University of Southern California.
McHugh, P. (1968). Defining the Situation: The
Organization of Meaning in Social Interaction.
Indianapolis, IN: Bobbs-Merrill.
Mead, (1932). The Philosophy of the Present. Chicago:
Open Court.
----- (1934). Mind. Self and Society. C.W. Morris (ed.).
Chicago: The University of Chicago Press.
Merton (1957). Social Theory and Social Structure.
Glencoe, IL: The Free Press.
Morley, D. (1980). The "Nationwide” Audience. London:
British Film Institute.
Nelkin, D. (1988). Communicating the risks and benefits
of technology. Presented to the 1988 Symposium on
Science Communication: Environmental and Health
Research. Annenberg School for Communication, Los
Angeles, CA.
Palmgreen, P. (1984). Uses and gratifications: A
theoretical perspective, in Bostrom, R. (ed.)
Communication Yearbook 8. Beverly Hills, CA: Sage.
Perry, R.W. (1988). The communication of disaster
warnings. Presented to the 1988 Symposium on Science
Communication: Environmental and Health Research.
Annenberg School for Communication, Los Angeles, CA.
Risen, J. and Johnston, O. (1990). "Major lending rate
cut by Fed" in the Los Angeles Times. 12-19-90, p. A-
1.
Robinson, J.P. (1981). Television and leisure time: A
new scenario. Journal of Communication 31:120-130.
22<
I
1
Rogers, E.M. (1988). Science communication about risk. I
Presented to the 1988 Symposium on Science !
Communication: Environmental and Health Research. !
Annenberg School for Communication, Los Angeles, CA.
----- and Dearing, J. (1988). Agenda-setting research: 1
Where has it been, where is it going? in J.A.
Anderson (ed.). Communication Yearbook 11. Newbury
Park, CA: Sage.
t
! Rokeach, M. (1973). The Nature of Human Values. NY: The
j Free Press. '
I ----- (1982). Values Survey. Palo Alto, CA: Consulting
Psychologists Press. ;
i i
' Sandman, P.M. (1988). Hazard versus outrage: The case of'
j radon. Presented to the 1988 Symposium on Science j
! Communication: Environmental and Health Research.
Annenberg School for Communication, Los Angeles, CA.
Schutz, A. (1962). On multiple realities, in Collected
Papers. I: The Problem of Social Reality. The
Hague: Martinus Nijhoff.
i
| Short, (1984). The social fabric at risk: Toward the
| social transformation of risk analysis. American
Sociological Review 49:711-725.
I
Simon, H. (1955). A behavioral model of rational choice.
Quarterly Journal of Economics 69:99-118.
----- (1978). Rationality as process and as product of
thought. Journal of the American Economic
j Association 68:1-16.
I
| Speisman, J., Lazarus, R.S., Mordkoff, A., and Davison, L.
j (1964). Experimental reduction of stress based on
| ego-defense theory. Journal of Abnormal and Social
i Psychology 68:367-380.
i
i Stiff, J., McCormack, M., Zook, E., Stein, T., and Henry,
R. (1990). Learning about AIDS and HIV transmission
: in college-age students. Communication Research
j 17:743-758.
i
| Turner, R.H., Nigg, J.M., and Paz, D.H. (1986). Waiting
for Disaster: Earthguake Watch in California.
! Berkeley, CA: University of California Press.
Waterman, D. and Grant, A. (1991). Cable television as an !
aftermarket. Journal of Broadcasting and Electronic I
Media 35(2):179-187.
Watzman, N. (1990). "Let'em Eat Pesticides: Taking the |
Worry out of Food" in The Nation. 12-10-90.
Withey, S. (1962). Reaction to uncertain threat, in <
Baker, G. and D. Chapman (eds.) Man and Society in j
Disaster. NY: Basic Books. I
I
Wober, J.M. (1978). Televised violence and paranoid !
perception: The view from Great Britain. Public j
Opinion Quarterly 42:315-321.
226
I
APPENDIX A
THE SAN BERNARDINO COMMUNITY SURVEY

1. We w ould lik e you to consider th e ways th a t you use fo u r m ass m e d ia in Y O U R d a lly life . In a ll of
th e ir vario us fo rm s. In clu d in g advertising, th in k about:
N ew spapers (d aily a n d w eekly) M ag a zin e s (local and n atio n al)
R a d io [FM and A M I T e le v is io n (netw ork, p u b lic, or cable)
Please ind icate how m u ch you rely upon these m edia by circlin g th e num ber th a t b est rep resen ts
how H e lp fu l th ey are to you for each item listed below .
In YOUR Dally Life. How
Helpful Are These Media To: n o t a t a l l e x t r e m e l y
H E LP FU L H E L P F U L
1 2 3 4 5
S ta y on to p o f w h a t Is h a p p e n in g is th e co m m u n ity'?
NEW SPAPERS............................................................... 1 2 3 4
R A D IO ................................................. 1 2 3 4
M A G A ZIN E S.................................................................................. 1 2 3 4
T E L E V IS IO N .................................................................................. 1 2 3 4
U n w in d a fte r a h a rd d ay o r w eek?
NEW SPAPERS.................................................... 2 3 4
R A D IO ...........- ..................................................... 2 3 4
M A G A ZIN E S........ .............................................. 2 3 4
T E L E V IS IO N ...................................................... 2 3 4
G a in in s ig h t in to w h y y o u do som e o f th e th in g s th a t y o u do?
N EW SPA P ER S ................................................... 2 3 4
R A D IO .................................................................. 2 3 4
M A G A ZIN E S.......................................................
2 3 4
T E L E V IS IO N ...................................................... 2 3 4
P ro v id e am u sin g th in g s to s h a re w ith o th e rs , such as Jokes a n d c arto o n s?
NEW SPAPERS........................................................................................ 1
R A D IO ........................................................................................................... 1
M A G A ZIN E S ................................................................................................ 1
T E L E V IS IO N ............................................................................................... 1
D is c o v e r b e tte r w ays to c o m m u n ic a te w ith o th e rs ?
N EW SPAPERS.............................................................................. 1
R A D IO ........................................... - ............................................................. 1
M A G A ZIN E S ................................................................................................ 1
T E L E V IS IO N ............................................................................................... 1
D ecid e w h e re to go fo r s e rv ic e s , such as h e a lth , fin a n c ia l o r h ou seh o ld ?
NEW SPAPERS............................................................................................ 1
R A D IO ............................................................................................................ 1
M A G A ZIN E S ................................................................................................ I
T E L E V IS IO N ................................................................................... 1
R e la x w h en y o u a re b y y o u rs e lf?
NEW SPAPERS............................................................................................ 1
R A D IO ............................................................................................................ 1
M A G A ZIN E S ................................................................................................ 1
T E L E V IS IO N ............................................................................................... 1
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
.4
4
4
4
4
4
S
s
5
5
5
5
S
5
5
5
5
5
5
S
5
5
VKjimu* utuioioi
In YOUR Dally Life. How
H elpful Are These Media To:
N O T A T A LL EXTREM ELY
H ELPFU L HELPFUL
1 2 3 4 5
F in d o u t h o w th e c o u n try U d oin g ?
NEW SPAPERS.......................................................................................... 1
R A D IO ........................................................................ 1
M A G A ZIN E S.............................................................................................. 1
T E L E V IS IO N ........................................................................... 1
Im a g in e w h a t y o u 'll b e lik e as y o u g row o ld er?
NEW SPAPERS.......................................................................................... I
R A D IO .......................................................................................... 1
M A G A ZIN E S .............................................................................. 1
T E L E V IS IO N , ...................................................... 1
G iv e y o u s o m e th in g to d o w ith y o u r frie n d s ?
NEW SPAPERS........................................................................................... I
R A D IO .......................................................................................................... 1
M A G A ZIN E S ..................................................................................... 1
T E L E V IS IO N ..................................... - ...................................................... 1
F ig u re o u t w h a t to b u y?
N EW SPAPERS........................................................................................... 1
R A D IO .......................................................................................................... 1
M A G A ZIN E S .............................................................................................. 1
T E L E V IS IO N .............................................................................................. I
T h in k a b o u t h o w to a c t w ith frie n d s , re la tir e s , o r p eo p le y o u w o rk w ith ?
NEW SPAPERS ............................................................ 1
R A D IO .......................................................................................................... I
M A G A ZIN E S ............................................................................................... 1
T E L E V IS IO N .............................................................................................. I
B a re fu n w ith fa m ily a n d M e n d s ?
NEW SPAPERS........................................................................................... I
R A D IO .......................................................................... 1
M A G A ZIN E S ................................................................ 1
T E L E V IS IO N ............................................................................................... 1
O b serve h o w o th e r* cope w ith p ro b lem s o r s itu a tio n s lik e y o u rs ?
NEW SPAPERS........................................................................................... 1
R A D IO ........................................................................................... .. 1
M A G A ZIN E S ............................................................................. 1
T E L E V IS IO N ............................................................................................. 1
K e e p u p w ith w o rld e v e n ts ?
NEW SPAPERS........................................................................................... 1
R A D IO ............................................................................................. 1
M A G A ZIN E S ............................................................................................... 1
T E L E V IS IO N ............................................................................................... 1
B e a p a rt o f e v e n ts th a t y o u e n jo y w ith o u t h a v in g to b e th e re ?
NEW SPAPERS........................................................................................... 1
R A D IO .......................................................................................................... 1
M A G A ZIN E S ............................................................................................... 1
T E L E V IS IO N ............................................................................................... 1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3 •
3
3
3
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
In YOUR Daily Life. How-
H elpful Are These Media To:
N O T A T A LL EXTR EM ELY
H E L P F U L H ELPFUL
1 2 3 4 5
G e t id e a * a b o u t h o w to ap p ro ach o th e r* in im p o rta n t o r d iffic u lt s itu a tio n *?
i NEW SPAPERS...............................................................................
j R A D IO ...............................................................................................
, M A G A ZIN E S ...................................................................................
, T E L E V IS IO N ........................................................... .....................
P la n w h e re to go fo r e v e n in g a n d w eeken d a c tiv itie s ?
NEW SPAPERS................................................. .
R A D IO .............................................................................................
I M A G A ZIN E S ...................................................................................
T E L E V IS IO N ..................................................................................
H a v e s o m e th in g to do w h en n o b o d y else la aro u n d ?
NEW SPAPERS...................................................................
R A D IO ...................................................................................
M A G A ZIN E S ......................................................................
T E L E V IS IO N ............................................................
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 S
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
!
2 . D o you subscrib e to c a b le te le v is io n ? (C heck one.)
YE S N O
3 . D o you w atch th e S a n B e rn a rd in o p u b lic access
c h a n n e l ( t 3 o n ca b le )? (C heck one.)
Agree
S t r ongl y
Disagree
S t r o n g l y
|4. D o you re g u la rly u se a V C R ? (C heck one.) ____
I
5 . D o you w atch p u b lic te le v is io n ? (C heck one.)___________________
I
6 . H ow m uch do Y O U ag ree w ith th e statem en t below .
i
I
It's h a rd e r th a n it used to b e to know w h at A m erica stands for. , 1 2 3 4
i
7 . Place a ch eck in th e space provided th a t best describes Y O U R w a y o f d ealin g w ith a ll o f th e c o n fu s in g
, aspects o f m o d em d ay life .
1 _____ I try to avoid th in k in g about th em .
| _____ I o n ly try to fig u re o u t those th in g s th a t are necessary to g et th ro u g h th e d ay.
1 _____ I try to fig u re th in g s o u t b u t I g en erally d o n t succeed.
j _____ I try to fig u re th in g s o u t and m ost o f th e tim e I succeed.
8 . T h in k a b o u t th in g s th a t cou ld th rea te n y o u r w ell-b ein g . H ow c o n c e rn e d are you th a t th e events listed
below co u ld h ap p en to Y O U o r to m em bers o f Y O U R fam ily? Please circle th e ap p ro p ria te n u m b er
fo r e a c h item .
A B O U T:
I A M E x tre m e ly
C oncerned
N ot A t A ll
Concerned
a. b ein g m ugged o r ro b b e d .................................................................
b . a n a tu ra l d is a s ter (e.g .. earth q u ake o r flood) th a t causes
dam age in m y a r e a ...........................................................................
c. b ein g u n a b le to affo rd th e risin g cost o f m edical c a re ........
d. a n u c le a r d is a s ter o r accid ent ...................................................
e. a n econom ic c ris is (e.g .. in fla tio n o r a serious re cessio n )..
f. g ettin g a serio us illn ess ..................................................................
g. fin d in g o u t th a t I live in an area th a t has
u n safe, w a te r, la n d o r a ir ..............................................................
h . losing m y Job ......................................................................................
3
3
3
3
3
3
3
4
4
4
4
4
4
4
231
9 . Now we w ould lik e to ask you a few questions abo ut how often you use vario u s m ass m edia.
a. A p proxim ately how m an y m in u te s a n d h o u rs do you spend reading m a g azin es each w eek?
H o u rs _____ M in u tes
b . Please lis t th e nam es o f an y m a g a zin e s th a t you rea lly look fo rw ard to re a d in g :
c. A p p roxim ately how m an y h ours o f te le v is io n do you w atch on an averag e d ay ?
W rite y o u r answ er h ere: ________
d . Please lis t an y te le v is io n c h a n n e ls (both cable and non-cable) th a t you re a lly lo o k
fo rw ard to w atching:
e. A p pro xim ately how m an y m in u te s a n d h o u rs do you spend read ing a n ew sp ap er each d ay?
H o urs _____ M in u te s
f. Please lis t th e nam es o f an y n ew sp apers th a t you re a lly lo o k fo rw ard to re a d in g :
g. A p pro xim ately how m an y h o u rs o f ra d io do you lis te n to on a n average d ay?
W rite y o u r answ er here: _______
h . Please lis t th e nam es o f an y ra d io s ta tio n s th a t you re a lly lo o k fo rw a rd to h e a rin g :
1 . Do you lis te n to p u b lic ra d io ? Yes No
10. W e w ould lik e yo u to consider th e n u m b e r o f d iffe re n t k in d s o f m e d ia th a t are availab le in the
C ity o f S an B e rn a rd in o . Please w rite a n u m b e r th a t corresponds w ith y o u r b e s t guess fo r each one.
N ew spapers th a t re p o rt „
lo cal events
Television chan nels th a t rep o rt
local events
F M an d A M rad io ch an nels
th a t can b e received
Cable television chan nels
Stores th a t re n t o r sell
videocassette (VCR) m ovies
M agazines th a t rep o rt on
local events
M agazines th a t re p o rt on
w o rld o r n a tio n a l events
P ublic television station s th a t
can be received
232
On the next page are 18 values listed in alphabetical order. Your task Is
to arrange them In order of their Importance to YOU, as guiding
principles in YOUR life. Each value is printed on a gummed label which
can be easily peeled off and pasted in the boxes on the left-hand side of
the page.
Study the list carefully and pick out the one value which is the most
important for you. Peel it off and paste it in Box 1 on the left.
Then pick out the value which is second most important for you. Peel it
off and paste it in Box 2. Then do the same for each of the remaining
values. The value which is least Important goes in Box 18.
Work slowly and think carefully. If you change your mind, feel free to
change your answers. The labels peel off easily and can be moved from
place to place. The end result should truly show how you really feel.
1 _________________
2 _______________________________________________
3 _______________________________________________
4 _______ _______________________________________
5 _______ _______________________________________
6 _______________
7 _______________________________________________
8 _______________________________________________
9 _______________________________________________
10______________________________________________ ;
11 ; ________________________
12 __________________________________
1 3 _______________________________________________
14 1
1 5 ______________________
1 6 ______________________
17 _____________
1 8 __________________
Whan you have flmjnoa. please proceed lo me next
page.
A C O M F O R T A B LE LIFE
(a p ro s p e ro u s lif e )
A N E X C IT IN G LIFE
(a s tim u la tin g , a c tlv a U fa )
A SENSE O F A C C O M P L IS H M E N T
( la s tin g c o n t r ib u t io n )
A W O R L D A T PE AC E
( Ir o o o f w a r e n d c o n f lic t)
A W O R L D O F B E A U T Y
( b e a u ty o f n a tu r e a n d t h e a r ts )
E Q U A L IT Y (b r o th e r h o o d ,
e q u a l o p p o r t u n ity f o r a ll)
F A M IL Y SECURITY
( ta k in g c a r e o f lo v e d o n e s )
FR EED O M
(in d e p e n d e n c e , f r e e c h o ic e )
HEALTH
(p h ysica l a n d m e n ta l w e ll- b e in g )
IN N E R H A R M O N Y
( fre e d o m f r o m in n e r c o n flic t)
M A T U R E LO V E
( s e x u a l a n d s p ir it u a l in tim a c y )
N A T IO N A L SECU RITY
(p ro te c tio n f r o m a t ta c k )
PLEASURE
(a n e n jo y a b le , le is u r e ly lif e )
S A L V A T IO N
(s a v e d , e te r n o l life )
SELF-RESPECT
(s e lf-e s te e m )
S O C IA L R E C O G N IT IO N
(re s p a c t, a d m ir a t io n )
TRUE FR IENDSHIP
(c lo s e c o m p a n io n s h ip )
W IS D O M
(a m a tu re u n d e rs ta n d in g o f life )
rank mese values in me same manner
as you did on me preceding page.
1_______________________
2_______________
■ 3 _______________________________________________
4 ______________________________________________
5
6 _______________________________________________
7 . _________________________________________________
8 ______________________
9 ____________________________________________________________
10______________________
n______
12 ___________________________________
1 3 ______________________________________________
1 4 ______________________________________________
1 5 ______________________
1 6 ___________
1 7 ___________________________________________________________________
1 8 ______________________________________________
Whan you nova flnuhea. Diaata proceed to ma nest
poga.
A M B IT IO U S
( h a r d * w o r liin g , a s p ir in g )
B R O A D M IN D E D
( o p o n -m ln d a d )
C A P A B LE
(c o m p o to n t, o H e c tiv a )
C L E A N
( n e a t, t id y )
C O U R A G E O U S
(s ta n d in g u p f o r y o u r b e lie fs )
F O R G IV IN G
( w illin g to p o r d o n o th e r s )
HELPFUL ( w o r k in g
f o r t h o w e lf a r e o f o th e r s )
H O N E S T
(s in c e re , t r u t h f u l)
IM A G IN A T IV E
( d a r in g , c r e a tiv e )
IN D E P E N D E N T
( s e lf - r e lia n t , s e lf- s u ffic ie n t)
IN TE LLEC TU A L
( In t e llig e n t , r e fle c t iv e )
L O G IC A L
( c o n s is te n t, r a tio n a l)
L O V IN G
( a ffe c tio n a te , te n d e r )
LOYAL
(fa ith fu l to o n e 's fn e n d s , g ro u p )
O B E D IE N T
( d u t if u l, r e s p e c tfu l)
PO LITE
( c o u rte o u s , w e ll- m o n n e r e d )
RESPONSIBLE
( d e p e n d a b le , r e lia b le )
SELF-CO NTR O LLED
( re s tr a in e d , s e lf- d is c ip lin e d )
235
You w ill recognize th e tw o lis ts o f values below . A s you lo o k th e n over a g a in , w e w o u ld lik e you to
th in k a b o u t w h a t fo rm e r P re s id e n t R eagan m e a n t w h en h e talk e d ab o u t re tu rn in g to " tra d itio n a l
A m e ric a n v a lu e s ." Please circle th e th re e values from each lis t th a t you th in k b es t rep resen t w hat
h e m e a n t.
L IS T O N E f Choose 3 ) L IS T TW O (Choose 3 )
A C o m fo rtab le L ife A m b itiou s
A n E x c itin g Life B roadm inded
A Sense o f A ccom plish m ent C apable
A W o rld a t Peace C lean
A W o rld o f B e a u ty C ourageous
E q u a lity Forgiving
F a m ily S e c u rity H e lp fu l
Freedo m H onest
H e a lth Im ag inative
In n e r H a rm o n y In d ep en den t
M a tu re Love In te lle c tu a l
N a tio n a l S e c u rity Logical
P leasure Loving
S a lvatio n Loyal
S elf-R esp ect O bed ien t
S o cial R ecog nition P o lite
T ru e F rie n d s h ip Responsible
W isdo m S elf-C o n trolled
N ow lo o k a g a in a t th e sam e lis ts and circle th re e values fro m each lis t th a t y o u th in k b est rep resent
w h a t P re s id e n t B u sh m ean s b y w an tin g to have "a k in d e r a n d g e n tle r n a tio n .".
L IS T O N E ( Choose 3 ) L IS T TW O (Choose 3 )
A C o m fo rtab le Life A m b itio u s
A n E x c itin g Life B roadm ind ed
A Sense o f A ccom plishm ent C apab le
A W o rld a t Peace C lean
A W o rld o f B e a u ty C ourageous
E q u a lity Forgiving
F a m ily S e c u rity H e lp fu l
F reedom H onest
H e a lth Im ag in ative
In n e r H a rm o n y In d epen den t
M a tu re Love In te lle c tu a l
N a tio n a l S e c u rity Logical
P leasu re Loving
S a lv a tio n Loyal
S elf-R esp ect O bed ient
S o cial R ecog nition P o lite
T ru e F rie n d s h ip Responsible
W isdo m S elf-C o n trolled
236
13. In Y O U R H O U S E H O LD :
a . W h a t la n g u a g e * are spoken regularly? C h eck as m an y as apply.
E n g lish S panish _ _ O th e r (please specify _ _ _ _ _
I b . D oes anyone lis te n to . w atch , or read S p an ish lan g uag e m e d ia ? Please ch eck one.
i
i Yes No
c. H o w m an y people age 1 6 o r o ve r live w ith you?
d . H o w m a n y people u n d e r th e age o f 16 live w ith you re g u la rly? ____ _
14. F in a lly , w e w o u ld lik e to kno w a little about you .
a . Y o u r b irth d ate: ______ ; _______
b . A re yo u em ployed? Please check one.
N o. a n d I am n o t looking fo r w o rk Yes, p a rt-tim e
________ N o. b u t 1 am looking fo r w o rk _ _ _ _ _ Yes. fu ll-tim e
c. H o w lo n g have you lived In th e S an B ern ard in o V alley a re a ? ____________ Y ears.
d . Y o u r sex: _ _ _ _ _ Fem ale _________ M ale
e. Y o u r occupation :
f. W h ic h o f th e follow in g best describes y o u r R acial o r E th n ic o rigin ? Please c h eck one.
A s la n _______C aucasian N a tiv e A m erican
B la c k _______H ispan ic
O th e r (Please specify__________ ; ___________________________ )
g. Please lis t a n y c o m m u n ity o r in te re s t groups w h ich you re a lly s u p p o rt, w h e th e r o r n o t y o u
a re a m e m b e r.
h . W h ic h o f these b ro ad categories b est describes y o u r to ta l h o u s e h o ld In c o m e fo r 1988?
j Please c h eck one.
' _____ U n d e r $ 1 0 ,0 0 0 $ 3 0 ,0 0 0 -6 3 9 .9 9 9 _____ $ 6 0 .0 0 0 -$ 6 9 ,9 9 9
$ 1 0 ,0 0 0 -6 1 9 .9 9 9 $ 4 0 ,0 0 0 -6 4 9 .9 9 9 $ 7 0 ,0 0 0 -6 7 9 .9 9 9
j _____ $ 2 0 ,0 0 0 -6 2 9 ,9 9 9 $ 5 0 ,0 0 0 -6 5 9 .9 9 9 _____ O ver $ 8 0 ,0 0 0
| 1. W h a t Is th e h ig h est level o f e d u c a tio n th a t you have com pleted? C irc le th e a p p ro p ria te n um ber.
G rad e school 1 2 3 4 5 6 7 8
H ig h school 1 2 3 4
C o lle g e /T e c h n ic al school 1 2 3 4 5
G ra d u a te school 1 2 3 4 5 6
T H A N K Y O U V E R Y M U C H F O R Y O U R C O O P E R A T IO N .
237
APPENDIX B
THE AUSTIN COMMUNITY SURVEY
1. We would like you to consider the ways that you use four mass media in YOUR
daily life. In all of iheir various forms, including advertising, think about:
Newspapers (daily and weekly)
Radio (FM and AM)
Magazines Oocal and national)
Television (network, public, or cable)
Please indicate how much you rely on these media by circling the number that best
represents how helpful they are to'you for each item listed below.
In Y O U R D aily Life. H ow
H elp fu l A re These M edia To:
N O T A T A LL
H ELPFU L
EXTREMELY
H ELPFU L
1
2 ..........
4
S
Stay on top of what is happening in the community?
NEWSPAPERS................................................. 1 2 3 4 5
RADIO............... ............................................... 1 2 3 4 5
MAGAZINES.................................................... 1 2 3 4 5
TELEVISION................................................... I 2 3 4 5
Unwind after a hard day or week?
NEWSPAPERS ________________________ 1 2 3 4 S
RADIO_____________ ___________ _______ 1 2 3
4 5
MAGAZINES___________________ ___ ___ 1 2 3 4 5
TELEVISION....... ........ ............................... I
G ain insight Into why you do some of the things that you do?
2 3 4 5
NEWSPAPERS______________ ______ ___ 1 2 3 4 5
RADIO_____________________ __________ 1 2 3 4 5
MAGAZINES__________________________ 1 2 3 4 5
TELEVISION__________________ _____1 2
Provide amusing things to share with others, such as jokes, cartoons, or
3 4 5
stories?
NEWSPAPERS. ___ ____ _____________ 1 2 3 4 5
RADIO_______________________________ 1 2 3 4 5
MAGAZINES___________ ____ _________ 1 2 3 4 5
TELEVISION__________________________ 1 2 3 4 5
Discover better ways to communicate with others?
NEWSPAPERS.__________________ _____ I 2 3 4 5
RADIO_____________________________ I 2 3 4 5
MAGAZINES..................................................... I 2 3 4 5
TELEVISION__________________________ 1 2
Decide where to go for services, such as health, financial, or household?
3 4 5
NEWSPAPERS..................... ........................... 1 2 3 4 5
RADIO —............. ........ ............ - ................ I 2 3
4 5
MAGAZINES............................ ........................ 1 2 3 4 5
TELEVISION..
238
In Y O U R D aily L ife , H o w N O T A T A L L
EXTREMELY
H e lp fu l A re T h e se M e d ia T o : H E L P F U L
H E L P F U L
l ._
1 1
....A. ......
S
Relax when you are by yourself?
NEWSPAPERS........................................... 1 2 3
4 5
RADIO......................................................... 1 2 3 4 5
MAGAZINES............................................. ....... 1 2 3 4 5
TEI-EVISION.............................................. 1 2 3 4 5
Find out how the country is doing?
NEWSPAPERS.......................................... ....... l 2 3 4 5
RADIO......................................................... 1
2
3
4 5
MAGAZINES..................................................... 1 2 3' .
4 5
TELEVISION............................................. l 2 3 4 5
Im agine what you'll be like as you grow older?
NEWSPAPERS.......................................... 2 3 4 5
RADIO......................................................... 1 2 3 4 5
MAGAZINES............................................. ........ 1 2 3 4 5
TELEVISION............................................. 1 2 3 4 5
Give you something to do with your friends?
NEWSPAPERS......................................... ........ 1 2 3 4 S
RADIO........................................................ ........ 1 2 3 4 5
MAGAZINES............................................ ......... 1 2 3 4 5
TELEVISION............................................ ........ t 2 3 4 5
Figure out what to buy?
NEWSPAPERS..............- ......................... 2 3 4 5
RADIO...................................................... 2 3
4 5
MAGAZINES............................................ 2 3 4 5
TELEVISION........................................... 1 2 3 4 5
Think about how to act with friends, relatives, or people you work with?
NEWSPAPERS......................................... 1 2 3 4 5
RADIO.______________ ________ ____ 2 3 4 5
MAGAZINES-......................................... 2 3 4 5
TELEVISION----- ------------------- ------
........... 1
2 3 4 5
Have fun with family and friends?
NEWSPAPERS_________________ _ _ .......- 1 2 3 4 5
RADIO_____________________ _____.. 1 2 3 4 5
MAGAZINES.............................................. 1 2 3 4 5
TELEVISION........... ......... ......... ............ ____ 1 2 3 4 5
Observe how others cope with problems or situations like yours?
NEWSPAPERS........ ............................... 1 2 3 4 5
RADIO_________ _____ - ............ ........... 1 2 3 4 5,
MAGAZINES............................ - ........... 2 3
4 5
TELEVISION............................................. - I 2 3 4 5
Keep up with world events?
NEWSPAPERS........................................ 2 3 4 3
RADIO........................................................ 1 2 3 4 3
MAGAZINES.......................................... 2 3 4 5
TELEVISION...................... .................... 2 3 4 5
240
In Y O U R D aily L ife, H ow
H elp fu l A re T hese M edia T o:
N O T A T A LL
H ELPFU L
I________ 2.____
Be a part of events that you enjoy without having to be
NEWSPAPERS...................................................
RADIO...................................:.............................
MAGAZINES......................................................
TELEVISION......................................................
Get ideas about how to approach others in important o
NEWSPAPERS...................................................
RADIO.................................................................
MAGAZINES......................................................
TELEVISION......................................................
Plan where to go for evening and weekend activities?
NEWSPAPERS...................................................
RADIO.................................................................
MAGAZINES......................................................
TELEVISION......................................................
H are som ething to do when nobody else is around?
NEWSPAPERS...................................................
RADIO........................................... ....
MAGAZINES.......................................................
TELEVISION.......................................................
lere?
2 3
2 ' 3
2 3
2 3
difficult situations?
2 3
2 3
2 3
2 3
EXTREMELY
HELPFUL
_i_________i .
YES NO
2. Do you7 (a) Regularly use a VCR? (Check one.) ____ ___
(b) Subscribe to C able Television ? (Check one.) ____ ___
(c) Watch Public television? (Check one.) ____ ___
(d) Own a C am corder? (Check one.) ____ ___
3. If you regularly use a VCR. how many times during an average week do you use the
VCR to: (If you do not regularly use a VCR. please skip to the next question.)
Record a program Ray a movie you've
for later playback?_______________ rented or purchased?_____________
Play an exercise or Play "home movies"
instructional videotape?__________ recorded on video?;______________
4. a . If you have a VCR, how long have you owned it?
Write your answer here. ___ ______________
b. Approximately how many videotapes do you have?
Prerecorded tapes Tapes you recorded yourself_______ Blank tapes_______
5. A new type of television called HDTV may be available soon. Before today, had you
heard about or read about HDTV? (Please check one)
Yes ____ No
6. In the space below, please list any pay cable channels (such as Showtime or The Disney
Channel) that you subscribe to.
2 4 1
7. We would now like 1 0 find out how helpful you think videocassette recorders, cable
television, and public television services are to YOUR daily life. If you do not use
one or more of these media, please indicate your best guess of how helpful they would be
to you if you did use them.
Please indicate how much you rely on these media by circling the number that best
represents how Helpful they are to you for each item listed below.
In Y O U R D aily L ife, H ow N O T A T A LL EXTREMELY
j
H elp fu l A re T hese M edia To: H ELPFU L H ELPFUL
1 2 3
4
4
Stay on top of what is happening in the community?
VCR................................................................... 1 2 3 4 5
CABLE TELEVISION....................................... 1 2 3 4 5
PUBLIC TELEVISION.................................... 1 2 3 4 5
Unwind after a hard day or week?
VCR................................................................... 1 2 3 4 5
CABLE TELEVISION....................................... 1 2 3 4 5
PUBLIC TELEVISION.................................... 1 2 3 4 5
G ain insight into why you do some of the things that you do?
VCR................................................................... 1 2 3 4 5
CABLE TELEVISION....................................... 1 2 3 4 5
PUBLIC TELEVISION.................................... 1 2 3 4 5
Provide amusing things to share with others, such as jokes, cartoons, or
stories?
VCR................................................................... 1 2 3 4 5
CABLE TELEVISION........................................ 1 2 3 4 5
PUBLIC TELEVISION.................................... 1 2 3 4 5
Discover better ways to comm unicate with others?
VCR.............. ................................................... 1 2 3 4 5
CABLE TELEVISION...................................... 1 2 3 4 5
PUBLIC TELEVISION.................................... I 2 3 4 5
Decide where t o go for services, such a s health, financial, or household?
VCR.................................................................. 1 2 3 4 5
CABLE TELEVISION....................................... 1 2 3 4 5
PUBLIC TELEVISION.................................... 1 2 3 4 5 ’
Relax when you are by yourself?
VCR............... ................................................... 1 2 3 4 5
CABLE TELEVISION.......................... ............ 1 2 3 4 5
PUBLIC TELEVISION.................................... 1 2 3 4 5 .
Find out how the country is doing?
VCR............. ..................................................... 1 2 3 4 5
CABLE TELEVISION........................................ I 2 3 4 5
PUBLIC TELEVISION.................................... 1 2 3 4 5
Imagine w hat you’ll be like as you grow older?
VCR................................................................... 1 2 3 4 5
CABLE TELEVISION........................................ 1 2 3 4 5
PUBLIC TELEVISION.................................... 1 2 3 4 5
In Y O U R D a ily L ife . H o w N O T A T A L L
H e lp f u l A re T h ese M e d ia T o : H E L P F U L
] _____2 ___L_
Give you something to do with your friends?
VCR................................................................... 1 2 3
CABLE TELEVISION....................................... 1 2 3
PUBLIC TELEVISION.................................... 1 2 3
Figure out what to buy?
VCR................................................................... 1 2 3
CABLE TELEVISION....................................... I 2 3
PUBLIC TELEVISION.................................... 1 2 3
Think about how to act with friends, relatives, or people you work with?
VCR................................................................... 1 2 3
CABLE TELEVISION....................................... I 2 3
PUBLIC TELEVISION ........................ 1 2 3
Have fun with family and friends?
VCR................................................................... 1 2 3
CABLE TELEVISION....................................... ! 2 3
PUBLIC TELEVISION.................................... I 2 3
Observe how others cope with problems or situations like yours?
VCR................................................................... 1 2 3
CABLE TELEVISION ............................ I 2 3
PUBLIC TELEVISION.................................... 1 2 3
Keep up with world events?
VCR .................................................. 1 2 3
CABLE TELEVISION....................................... 1 2 3
PUBLIC TELEVISION.................................... 1 2 3
Be a p art of events that you enjoy without having to be there?
VCR ........................ ........................... 1 2 3
CABLE TELEVISION....................................... 1 2 3
PUBLIC TELEVISION .............................. 1 2 3
G et ideas about how to approach others in important or difficult situations?
VCR................................................................... I 2 3
CABLE TELEVISION....................................... 1 2 3
PUBLIC TELEVISION................................ 1 2 3
Plan where to go for evening and weekend activities?
VCR.................................................................. 1 2 3
CABLE TELEVISION....................................... 1 2 3
PUBLIC TELEVISION.................................... I 2 3
Have som ething to do when nobody else is around?
VCR................................................................... 1 2 3
CABLE TELEVISION....................................... 1 2 3
PUBLIC TELEVISION.................................... 1 2 3
EXTREMELY
H ELPFU L
A_____ L
5
5 .
5
t A 1 A V A L A LA
8. H ow m uch do YO U a g re e w ith the statem ent below.
Agree D isagree
S ia r x h :_____________________________ Strongly.
It's harder than it used to be to know
w hat A m erica stands for. 1 2 3 4 5
9. Place a check in the space provided that best describes YO UR way o f dealing with all of
the c o n fu sin g asp ects o f m odem day life.
1 try to avoid thinking about them .
1 only try to figure out those things that are necessary to gel through the day.
I try to figure things out, but I generally don't succeed.
1 try to figure things out and m ost o f the time I succeed.
10. T hink about the things that could threaten your well-being. How co n cern ed are you
that the events listed below could happen to YOU o r to m em bers o f Y O UR family? Please
circle the appropriate nu m b er for eac h item.
I AM — Extrem ely
A B O U T : Concerned
. a. being m ugged o r robbed......................................... 1
b. a natural disaster (e.g., earthquake o r flood)
. that cau ses dam age in m y area............................ 1
c. b ein g unable to afTord the rising co st o f
m edical care.................................................................. 1
d. a nuclear disaster o r accident.............................. 1
; c. an econom ic crisis (e.g., inflation o r a
serious recession).......................................................... 1
f. getting a serious illness.......................................... I
g. finding out that 1 live in an area that has
unsafe w ater, land, o r air........................................ 1
h. losing m y jo b ................................................................. 1
■11. W e w ould like you to co n sid er th e num ber o f d ifferen t k in d s o f r
available in the C ity o f A u stin . Please w rite a n u m b e r that corresponds with your
b e st guess for each one.
N ew spapers th at report- ________ Television channels that report
local events local events
FM and A M radio channels ________ Cable television channels
that can be received
Stores that rent o r sell ________ M agazines that report on
videocassette (V C R ) m ovies local events
_______ M agazines that report o n ________ Public television stations
world o r national events that can be received
12. In a typical week, how m any television program s do you watch in which a doctor or
o th er health professional is likely to appear? (The program s can include sitcom s,
dram as, soap operas, new s show s, talk show s, and other kinds o f program s.)
W rite your answ er here_________
Not At All
C oncerned
2 3 4 5
2 3 4 5
2 3 4 5
2 3 4 5
2 3 4 5
2 3 4 5
2 3 4 5
2 3 4 5
nedia that arc
13. How many times in the past year have you talked to your doctor about a health topic or
concern that you've heard about through the media?
Write your answer hcre________
244
]4. Now we would like 10 ask you a few questions about how often you use various m ass
media.
a. Approxim ately how m any m in u tes and h o u rs do you spend reading
m agazines each w eek?
_____________ Hours Minutes
b. Please list the names o f any m agazines that you really look fo rw a rd to re ad in g :
c. Approxim ately how m any hours o f television do you watch on an av erag e day?
W rite your answ er here:.
d. Approxim ately how m any hours of television did you watch y e sterd a y ?
W rite your answ er here:
e. Please list any television ch an n els (both cable and non-cable) that you really look
fo rw a rd tb w atching:
f. A pproxim ately how m any m in u tes an d h o u rs do you spend reading a n e w sp a p e r
each day ?
______________ Hours M inutes
g. Please list the nam es o f any n ew sp ap ers that you really look fo rw a rd to re a d in g :
h. A pproxim ately how m any hours o f ra d io do you listen to on an a v e ra g e day?
W rite your answ er here:.
i. A pproxim ately how m any hours o f ra d io did you listen to y e sterd a y ?
W rite your answ er here:.
j. Please list the nam es o f any ra d io sta tio n s that you really look fo rw a rd to
h earin g .
k. Do you listen to p u b lic ra d io ? (Please check one) Yes No
245
15. In YOUR HOUSEHOLD: •
a . What languages are spoken regularly? { Check as many as apply.)
English Spanish Other (please specify)______________________
b. D oes anyone listen to, watch, o r read Spanish language media? (Check one.)
■ Yes __________No
c. How many people age 16 o r over live with you?____________
d. How many people u n d er Ihe age of 16 live with you regularly?__________
16. Finally, we would like to know a little about you.
a . Your birth d ate:_________________________.
b. Are you employed? (Please check one.)
No, and 1 am not looking for work Yes, Part-time
No, but I am looking for work Yes, Fuli-time
c. How long have you lived in the Austin area?____________Years.
d. Your sex: Female Male
e. Your occupation:__________________________________ ________________
f. Which o f the following best describes your Racial or Ethnic origin?
Please check one.
Asian Caucasian _______ Native American
Black Hispanic
Other (Please specify _________________________ )
g. W hich o f these broad categories best describes your total household income for
1988? (Please check one.)
Under J 10.000_____________ $30.000-S39,999 _____ S60,000-$69.999
510,000-819,999___________ $40,000-549.999 $70,000-579.999
$20,000-529.999 _____ $50,000-559.999 _____ Over $80,000
h. What is Ihe highest level o f education that you have completed? Circle the
appropriate number.
Grade school 12 3 4 5 6 7 8
High school 1 2 3 4
College 1 2 3 4 5
Graduate school 1 2 3 4 5 6
TH A N K YOU VERY M U CH FO R COM PLETING THIS FO RM ! To return it. simply fold
it and place it in the postage paid envelope we have provided. Seal the envelope and drop it
in any mailbox.
If you have any questions about this survey, please call Prof. August Grant at 471-4071,
or write to us at:
August Grant, Assistant Professor
Department of Radio/Television/Film
C M A 6.I18
The University of Texas at Austin
Austin, Texas 78712

Asset Metadata

Creator Loges, William E. (author)

Core Title Canaries in the coal mine: Perceptions of threat and media system dependency relations.

Contributor Digitized by ProQuest (provenance)

Degree Doctor of Philosophy

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

Tag mass communications,OAI-PMH Harvest

Language English

Advisor Ball-Rokeach, Sandra (committee chair), [illegible] (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c17-726028

Unique identifier UC11344546

Identifier DP22477.pdf (filename),usctheses-c17-726028 (legacy record id)

Legacy Identifier DP22477.pdf

Dmrecord 726028

Document Type Dissertation

Rights Loges, William E.

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