MATHEMATICAL MODELS OF COMMUNICATION AND <  :
BELIEF CHANGE PROPORTIONAL CHANCE} ‘2 13;“;2 '
ACCUMULATED INFORMATION, AND BELIEECERTAINTY :g- ’

. Dissertation for theDegree of FIND; -' ’
MICHIGAN STATE UNIVERSITY
» JEFFREY E DANES ‘ .
1976

Rena-”Walt! 9.33:; W

I; if" 8 l2: R Y

Michigan State
University

    
     

E.

  

This is to certify that the

thesis entitled ‘

MATHEMATICAL MODELS OF COMMUNICATION
AND BELIEF CHANGE: PROPORTIONAL CHANGE,
ACCUMUIATED INFORMATION, AND BELIEF CERTAINTY

presented by

Jeffrey E. Danes

has been accepted towards fulfillment
of the requirements for

Ph-D- degree in _Comrmmica_tion

 

\

l ’ \. , N.
\a’v ' ‘h I («J r 4.» BN(

./ . I V
\\__ Major professor I

Date July 29, 1976

 

0-7639

 

 

 

ABSTRACT
MATHEMATICAL MODELS OF COMMUNICATION

AND BELIEF CHANGE: PROPORTIONAL CHANGE,
ACCUMULATED INFORMATION, AND BELIEF CERTAINTY

BY

Jeffrey B. Danes

Three models of communication and belief change were
proposed and tested. The proportional change model stated
that receivers change their beliefs in the direction of the
message with the resulting belief change being proportional
to the amount of change requested. The accumulated infor-
mation model was based upon the same logic; however, it
stated that belief change would be inhibited by the degree
to which information has been accumulated into the belief.
It was hypothesized that accumulated information and belief
certainty would be positively correlated; hence, a belief
certainty model was also proposed and tested.

The three models were tested with two sets of data,
and the results obtained showed clear support for the accu—
mulated information model. The belief certainty model was
the most inferior of the three. Although a positive corre-
lation was found between accumulated information and belief
certainty, the "informed neutrals" were almost nonexistent;

while "uninformed resolutes" prevailed. Regardless of

 

Jeffrey E. Danes
initial belief, the "uninformed" were the most affected by

the belief-change messages.

MATHEMATICAL MODELS OF COMMUNICATION
AND BELIEF CHANGE: PROPORTIONAL CHANGE,

ACCUMULATED INFORMATION, AND BELIEF CERTAINTY

BY

Jeffrey E: Danes

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Communication

1976

Accepted by the faculty of the Department of
Communication, College of Communication Arts, Michigan
State University, in partial fulfillment of the require-

ments for the Doctor of Philosophy degree.

€4.le

Diiéctor of Dissertation

 

Guidance Committee: ‘>§§11LLCMQ'1 , Chairman

LL /\ ( b(,_1_// l((( ~\ Q

[1% a L (:74 // - Fr

L’V'N.

 

 

 

 

All of this work and more is dedicated to the three
most important people in my life: Katherine, Nathan, and

Eric Danes.

ii

ACKNOWLEDGMENTS

I would like to express my thanks and appreciation to
the two persons who were most influential during my gradu-
ate training in Communication: Dr. Joseph Woelfel and
Dr. John Hunter. I thank Dr. Woelfel for the many stimula-
ting conversations and the many hours put into this disser-
tation. His help was deeply appreciated. I thank Dr.
Hunter, for without his help, this dissertation might never
have been completed. His assistance in math modeling and
in the analysis of model data was very much appreciated.

My appreciation also goes to Drs. Bettinghaus, Atkin, and
Ebel. Dr. Bettinghaus provided many useful and perceptive
comments on earlier drafts of this dissertation. Both Drs.
Atkin and Ebel also did much to bring the dissertation into
a clearer focus. Dr. Atkin gets a special word of thanks
for allowing me time to write when I was suppose to be in
the classroom.

My thanks and gratitude are also given to many of my
colleagues and students: I express my thanks to Nick
Stoyanoff for the many hours spent listening and criticizing
the ideas presented in this dissertation. Nick also de-
serves thanks for his help during data collection. Patti

Kluck and Donna Borris who also helped collect the data are

iii

also thanked for their efforts. Paul Taperek and Helen

Short gave continual support and overall encouragement;

for this, I express my thanks and appreciation to them.
Lastly and most importantly, I would like to thank

my family: Katherine, Nathan, and Eric, for their patience

and love while I spent three years working toward the Ph.D.

in Communication.

iv

Chapter

I

II

III

IV

TABLE OF CONTENTS

INTRODUCTION . . . .
Introduction to the Problem
The Proportional Change Model
Accumulated Information .
The Accumulated Informational Model
Belief Certainty . .
The Belief Certainty Model
METHOD . . . . .
Pilot Study . . . .
Accumulated Information
Belief Topic and Message
Selection . . . .
Testing the Models: Main Study
Experimental Design .
Procedures . .
Measurement . . .
PRELIMINARY ANALYSES . .

Scale Construction,
and Reliability .

Message Effect Manipulation Check

RESULTS . . . .

Testing the Models

Accumulated Information and

Belief Certainty

Evaluation,

Page

wNKDQAbUJ

HF4

16

16

17

18

18

20

22

22

24

27

27

32

Chapter

V DISCUSSION . . . . . . . .
APPENDICES . . . . . . . . . .
A. Belief topics and standard deviations

REFERENCES

for the belief certainty and accumulated
information measures used in the pilot

study . . . . . . . .

The military message used in the ex-
periment. The material was adapted

from an actual news story presented in
the March 8, 1976 issue of Time: "That

 

Alarming Soviet Buildup." . . .

The nuclear message used in the experi-

ment. The material was adapted from
an actual news story presented in the
March 8, 1976 issue of Time: "The

Struggle Over Nuclear Power." . .

The pretest questionnaire. Adminis—
tered on May 13, 1976 . . . .

Posttest questionnaire. Administered
May 13, 1976. . . . . .

Delayed posttest. Administered May
20, 1976. . . . . . .

Multiple groups, oblique cluster
analysis (key to indicators). .

vi

Page

36

40

40

43

47

51

60

68

76

78

Table

1.

8a.

8b.

8c.

LIST OF TABLES

Reliability Estimates for the Pretest
Measures of Belief and Accumulated Information

Mean Pretest, Posttest, and Change Values
and Standard Deviations for the Experimental
and Control Situations . . . . . .

ANOVA Regression Summary for Effects of
Nuclear Message upon Nuclear Belief Change .

ANOVA Regression Summary for Effects of
Military Message upon Military Belief Change

Correlations Between Ab and d for the Nuclear
and Military Topics in Both the Experimental
and Control Situations . . . . . .

Belief Change Means and Sample Size for the
Three Levels of Accumulated Information and
the Three Levels of Belief for the Nuclear
Message . . . . . . . . . .

Belief Change Means and Sample Size for the
Three Levels of Accumulated Information and
the Three Levels of Belief for the Military
Message . . . . . . . . . .

Number of Subjects in Each of the Cells for
the Three Levels of Belief and the Three
Levels of Accumulated Information for the
Nuclear Pretest Group . . . . . .

Number of Subjects in Each of the Cells for
the Three Levels of Belief and the Three
Levels of Accumulated Information for the
Military Pretest Group . . . . . .

Number of Subjects in Each of the Belief-
Accumulated Information Cells for Both the
Nuclear and Military Topics . . . . .

vii

Page

24

24

25

26

27

30

31

33

34

34

Figure

LIST OF FIGURES

Parametric curves for the prOportional
change model with a parameterized. The
curves graphically represent predicted
change as a function of the amount of
change requested . . . . . . .

Parametric curves for the accumulated
information model with a held constant

at one; and accumulated information (i )
controlled. The curves graphically
represent predicted change as a function
of the amount of change requested, in-
hibited by prior information accumulation .

Parametric curves for the belief

certainty model with a controlled. The
curves graphically represent predicted
change as a function of the amount of
change requested, inhibited by belief
certainty . . . . . . . . .

Regression curves of nuclear belief
change on initial belief with accumulated
information parameterized . . . . .

Regression curves of military belief

change on initial belief with accumulated
information parameterized . . . . .

viii

Page

10

14

30

32

CHAPTER I

INTRODUCTION

In much of communication research, a major construct
used for the assessment of communication effects is typi-
cally "attitude" change. It is recognized by numerous com-
munication scholars, however, that communication affects a
multitude of other mental variables. That is, not only are
"attitudes" possibly altered by the reception of messages,
but "knowledge," "perception," "opinions," "meaning," be—
lief," and a host of other constructs may change during the
process of communicating.

A current view adopted by many communication scholars
is that communication has its primary effect upon the ways
in which the human structures, maps, or organizes the sym—
bolic environment so that it becomes understandable
(Roberts, 1971). The structuring and restructuring of one's
symbolic environment goes by many different terms; however,
the term currently in vogue for much of communication re-
search is Boulding's (1956) construct of the "image."

Boulding (1956) has conceptualized the image to be
what one believes to be Eruef-one's subjective knowledge of
the various aspects of the physical, social, and symbolic
world. By another name, the image construct has also been

1

2
referred to as belief; Bem (1970) has offered the following
treatment of belief:

If a man perceives some relationship between
two things or between something and a char-
acteristic of it, he is said to hold a belief.
For example, he might suppose asteroids to be
round, the dean of women to be square, God

to be dead, men to love freedom, himself to
dislike spinach, and Republicans to promote
congress. Collectively, a man's beliefs
compose his understanding of himself and his
environment (pp. 4-5).

The focus of this study is upon the change of beliefs
using messages; the belief of concern is what Rokeach (1968)
has termed the "authority" belief and Fishbein and Azjen
(1975) "informational" beliefs.

Many of our beliefs are formed neither on the
basis of direct experience with the object of
belief nor by way of some inference process.
Instead, we often accept information about

some object provided by an outside source.

Such sources include newspapers, books, mag-
azines, radio and television, lectures, friends,
relatives, coworkers, etc. Beliefs formed by
accepting the information provided by an out-
side source may be termed informational beliefs
(Fishbein & Ajzen, 1975, p. 133).

 

For the present, belief is conceptualized as the con-
ceived truth value of an assertion or statement. And as
such, a belief exists when one thinks "true," "uncertain,"
or "false" to a given assertion or statement. Gradations
of truth value may also be considered; thus, belief is

viewed as continuum ranging from "true" to "false."

Introduction to the Problem

 

Currently there is an increased interest in the modeling
of communication effects within the domain of "passive com-
munication." Passive communication as opposed to active
communication restricts the domain of inquiry to those situ-
ations where a receiver decodes a message sent by a source,
and the receiver ". . . is not asked to respond to the mes-
sage in any active way (argue, give money, risk life, etc.)
or openly commit himself to a position that might be contrary
to that taken by his peers or reference groups" (Hunter and
Cohen, 1972, pp. 4-5). This context for which to study com-
munication effects is desirable in that it does represent
many forms of symbol transmission actually applied to prac-
tice: many forms of the mass communication of news and many
forms of advertising fall within the domain of passive com-
munication.

Within the passive communication context, the simplest
possible model for the effects of communication upon belief
change is one that predicts changes that are equal to the
amount of change advocated. Of the identical genre, a
slightly more complex model would predict changes that are
proportional to the amount of change advocated. The propor-
tional change model is derived from the equal change model

below.

4

The Proportional Change Model

 

With the assumption that a message is perfectly effec-

 

tive, a change in belief (Ab) as a result of receiving a

belief—change message would result in maximum change; i.e.,
all of the change requested would be obtained. The perfect
effects model is quantified below (for the purpose of quan-

tification, true will be set to zero and false to 100):

Ab = -b0 (1)
where, Ab = the change in the original belief
b0 = the original belief

bO = the subjective probability for
truth—value; totally true equals
0 and totally false equals 100.

Consistent with the assumption, this model states that a
perfectly effective message would produce a change equal to
the negative of the original belief (b0); this would render

the next belief (bl) to be zero (maximum truth):

bl = b0 + Ab (2)
b1 = bo ’ b0 (3)
bl = o (4)

This model, however, assumes that the message is per-
fectly effective; as is well known among communication
scholars, the impact of a message typically yields changes
that are proportional to those recommended.

The proportional change model was first suggested by
French (1956) and has since been elaborated by a variety of

authors (Hovland and Pritzker, 1957; Anderson and Hovland,

5
1957; Anderson, 1959; Anderson, 1965; Hunter and Cohen,
1972; Anderson, 1971; Whittaker, 1967). This model asserts
that the change in a belief is proportional to the discrep-
ancy between the belief and the message. If bO is the

initial belief and the message asserts the truth of the

statement, then this model becomes (see Figure 1.):

 

 

Ab = -abo (5)
where, a = a parameter for measured
message effectiveness
a = Oidil
lA
— i> b0
d = 1/4
Vb d = 1/2
a = l

 

Figure l. Parametric curves for the proportional
change modelMdiiuxparameterized. The curves
graphically represent predicted change as a
function of the amount of change requested.

If it were absolutely true that the amount of belief
change obtained was proportional to the amount of change
commanded, then once a was measured, the future states of
b (i.e., bn) could be clearly and unambiguously predicted

0
by the calculation of the prediction equation latent within

6

this difference model. The change in b (Ab) may also be

0
written as:
Ab = bl - b0 (6)
where, bl = the next belief
b0 = the original belief
Adding b0 to both sides of equation (6) gives:
Ab + b0 = b1 (7)
= +
b1 b0 Ab (8)

Equation (5), however, specifies that Ab is equal to -db0;

thus:

bl = b0 - abo (9)

b = (l - d)b0 (10)

As such, the model predicts that after one exposure to a
belief-change message, the next belief ((bl) will be (1 - a)
times the original (last) belief (b0). Thus, the general
predictive equation for n repeated exposures to a change

message becomes:

11

b = (1 - a) b

n 0 (11)

The resulting predictive equation for n message repetitions
of one change message results in an exponentially declining
sequence, indicating that as n grows large, the acceptance

of the truth claim increases; that is as n_+ w, bn + 0. As

such, it is clear that even with relatively minute changes,

the accumulative effects of message repetition may be rather

dramatic.

There are a number of communication variables that are
believed to be related to the accuracy of predicting belief
change. Given the initial assumption that belief change
will be proportional to the amount of change commanded, in-
creased accuracy of prediction would be enhanced if one in-
corporated those communication variables that impede, enhance,
and interact with the amount of change commanded. That is,
if one knew the values for the credibility of source, message
sidedness, message distraction, evidence, topic relevance,
topic interest, the entertainment value of the message, ac—
cumulated information, belief certainty, etc. then one might
be in a better position to accurately predict the amount of
change expected. This dissertation focused upon the follow-

ing variables: Accumulated information and belief certainty.

 

 

Accumulated Information

 

For some years we have known that "established" atti-
tudes are more difficult to change than are "de novo" atti-
tudes (Hovland, 1959). Anderson (1959, 1965) and Rosenberg
(1968) suggested that this effect could be accounted for
within the context of the discrepancy model if the parameter
a were to decrease as a function of accumulated information.
This model was specialized by Saltiel and Woelfel (1975) who

asserted that the parameter a is

n n (12)

8
where n is the number of messages ever received on the topic.
If a were to decline in this manner, then the belief after 3
messages would simply be the arithmetic mean of those mes-
sage values.

Using the average number of times an individual has com-
municated with his or her significant others about a partic-
ular topic (a composite of American values) as a measure of
accumulated information, Saltiel and Woelfel (1975) have
provided path analytic support for this hypothesized rela-
tionship. Saltiel and Woelfel (1975), however, failed to
present a conceptual definition for "information" or for
"accumulated information"; nonetheless, they do assert that
messages carry information and that the reception of messages
causes information to internally accumulate within the re-
ceiver. On the definition of information for human commun-
ication, Lin (1973) has argued that information exists only
when a receiver is familiar with the symbols of a message:
unfamiliar symbols convey little to no information.

Further, Lin (1973) has argued that information may be

 

". . . defined as a set of symbols which both the source

and receiver are familiar" (p. 23). As such, accumulated

 

information may be defined as the mental aggregation (stor—

 

age) of a set of familiar symbols sent from a source to a
receiver. Accumulated information is seen to differ from
"knowledge" in that knowledge implies "correctness"; where-

as the notion of accumulated information makes no such

-implication.

The relationship of accumulated information to belief
change messages may be interpreted in the following way:
when a receiver decodes a message advocating belief change,
the receiver according to the proportional change model,
makes a mental comparison between his or her initial be-
lief and the proposed belief, and then yields proportionately.
Likewise, other mental comparisons are likely; the accumu-
lated information hypothesis implies that a receiver not
only makes belief comparisons but also assesses the degree
to which he or she is "informed" about the belief topic.

If one is ngt_informed; that is, if one cannot retrieve
prior message content (pro or con) then this new information
compared to the old (none) takes precedent and consequently
alter the initial belief. Further, if one has accumulated
much information, then during the comparison process this
information might be retrieved and used in defense of the
initial belief, resulting in little to no belief change.

A model which incorporates the accumulated information
hypothesis into the proportional change model is presented

below.

The Accumulated Informational Model

 

According to the accumulated information hypothesis,
the amount of information that one has accumulated into a
belief will be inversely related to that belief's suscepti-

bility to change. In the model presented below, the

10
accumulated information hypothesis is compactly incorporated

into the proportional change model (see Figure 2):

 

_ _ 1
Ab - Odm) b0 (13)
0
b0
Ab = “0(i—fi) (l4)
0
where, i0 = the amount of information that
has been accumulated into the
original belief (b0).
1\
> be
Ab io= 2
10: l
10: 1/2
i0= 0

 

Figure 2. Parametric curves for the accumulated
information model with a held constant
at one; and accumulated information (i0)
controlled. The curves graphically
represent predicted change as a function
of the amount of change requested, in-
hibited by prior information accumulation.

To derive the predictive form of the above change
equation, it is first noted that a change in belief may also

be written as:

Ab

ll
0‘
1
O‘

(15)

11

Adding b to both sides of the equation gives:

0
Ab + b0 = b1 (16)
bl = Ab + b0 (17)
bl = b0 + Ab (18)

Substituting the identity given in (13) above yields:

.39.,
io+l

b = b - d(

1 o (19)

Factoring gives:

a
bl — [1 — (1+i0)] b0 (20)

 

The value for 10 after the reception of one belief
message may or may not change. If accumulated information
does not change, then the model presented below is appro-

priate for predicting the number of messages necessary to

change the belief to a desired value:

n
__ _ OI
bn ‘ L (1+1 )] b0 (21)

 

0

If, however, accumulated information does change as a re-
sult of message repetition, then the succeeding beliefs may

be predicted by the following equations:

 

 

 

 

_ r _ a
b = F1 - (—9‘——)- - ( 0‘ ) b (23)
2 l+il 1+1O 0

 

_ H _ d
102 ‘ b0 k=0 [1 (1+1 )] (24)

12

As such, belief after the nth message repetition is:

n-l

_ H -___L 25
bn _ b0 k=0 [1 (1+ik)] ( )

Belief Certainty

 

The forthcoming argument makes the following claim:
accumulated information and belief certainty are positively
correlated. In ordinary English "don't know" and "uncertain"
sometimes mean the same thing. If one does not "know" or is
not informed about a given belief issue, then it might be
presumptious for that person to assume an extreme belief
stance. On the other hand, if one does assume an extreme
belief stance, then it is likely that one is also informed
about the topic of concern, an assumption implied by the
work of Patterson and McClue (1976).

However, if the relationship between accumulated infor-
mation to belief certainty were perfect then there would be
no people of the following kind: those who are well informed
about a given belief topic but refuse for one reason or
another to take an extreme stance (the "informed neutrals"),
and those who are not informed but yet adopt an extreme be—
lief stance (the "uniformed resolutes").

For many years investigators have believed that people
with extreme beliefs are more resistant to change than those
with more neutral beliefs (Brim, 1955). This principle was
dubbed the "polarity" principle by Osgood and Tannenbaum

(1955) when they incorporated it into their congruity theory.

13
The polarity principle was used in conjunction with discrep-
ancy theory in a version of "information processing" theory
by Hunter and Cohen (1972) and the model below is adapted

from theirs.

The Belief Certainty Model

 

Like that of accumulated information, it was hypothe-
sized above that the certainty of a belief should also be
inversely related to change; as such, the model for belief

certainty is:

_ _ 1

Ab _ (17%;) b0 ‘26)
b0

Ab = " (Titer) ‘27)

0
Where, c0 = the certainty to which the belief is
held
B = a scaling parameter

Belief is operationalized as a response to a subjective
probability scale such that 0 represents totally "true" and
100 represents totally "false." The uncertain or uncom-
mitted position is represented by a score of 50. As the be-

lief departs from the uncertain (50) position, then Bc may

0

be defined as:

BC0 = 2 b0 - 50 (28)

And, the parametric curves for the belief certainty model

are presented in Figure 3.

l4

 

Ab

 

 

Figure 3. Parametric curves for the belief certainty
model with a controlled. The curves graph—
ically represent predicted change as a
function of the amount of change requested,
inhibited by belief certainty.

The predictive form of this model is:

k=0 k

The validity of the predictive equations for the pro—

n-l ~ u
bn = 130 H [1 " (TIE—Ff] (29)

portional change, accumulated information, and the belief
certainty models have their foundations upon the validity

of the following change equations:

Ab = - abo (the proportional change model)
b
Ab = - a(ifT—J (the accumulated information
+1
0 model)
b0
Ab = — a<IIBE—) (the belief certainty model)

0

15
Thus, the change equation forms of these models are tested

below.

CHAPTER II

METHOD

This chapter discusses a pilot study that was designed
to evaluate two methods of measuring accumulated information
and to select the belief topics and messages to be used in
the main study. The main study begins with a discussion of
the experimental design, and next the experimental procedures
are presented. This chapter concludes with a discussion of

the measurement instruments used in the main study.

Pilot Study

 

A questionnaire was administered to 33 introductory Com-
munication students at Michigan State University with the
purpose of (a) evaluating two methods for the measurement of
accumulated information, and (2) selecting belief tOpics and
messages to be used in the main study. The 18 belief topics
used for the pilot study appear in Appendix A.

Accumulated Information - With an example topic, below

 

are the scales used to measure accumulated information in

the pilot study:

About the relative strength Q: the U.S.A. and the
U.S.S.R. military forces, HOW MANY TIMES have you
received communications on this topic?

 

 

 

16

I l

 

17

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7) eight(8) nine(9)
ten(10) 11-25 26-50 51-75 76-100

more than 100

On this topic, I know

Nothing : : : : : : : : A lot

 

The exposure scale was scored as indicated by the
values in parentheses, and 11-25 = 18, 26-50 = 38, 51-75
= 63, 76-100 = 88, and 100+ was assigned 112. The Likert-
type informed scales were scored from 0 to 6 with six indi-
cating "A lot."

It was expected that the exposure scales would be
highly correlated with the informed scales, but the result-
ing correlation averaged over the 18 topics equaled only
.45. However, when the exposure scales were transformed
such that the transformed value = log (x + .5), the corre-
lation averaged over the 18 topic equaled .70. A correla-
tion of sufficient magnitude to justify the use of both
scales for the measurement of accumulated information--when
the exposure scale is logrithmically transformed.

Belief Topic and Message Selection - Belief was measured

 

using a bipolar scale with "true" set to 0 and "false" to
100; belief scores were obtained by subtracting the reported
value from 100 (an example of this scale appears in Appendix

D).

18

As noted earlier, a fair test and comparison of the
three models of communication and belief change demand that
the beliefs used exhibit variance in both belief certainty
and in accumulated information. The resulting standard de-
viations for the 18 beliefs and 36 information measures
appear in Appendix A. All showed suitable distributions;
thus, two of the more general topics were selected to be
used in the main study: (a) The nuclear production of elec—
tricity is potentially more dangerous than the conventional
methods of producing electricity, and (b) The U.S.S.R.
military forces are becoming superior to the military forces
of the U.S.A. Hereafter, these two statements will be
respectively referred to as the "nuclear" and the "military"
beliefs.

Each of the 18 topics that appear in Appendix A also
appeared in a variety of current news magazines as news
stories. Those stories corresponding to the nuclear and
military beliefs were retrieved, modified, and transformed
into the experimental messages used in the main study.

These messages appear in Appendices B and C, and they will

hereafter be called the "nuclear" and the "military" messages.

Testing the Models: Main Study

 

Experimental Design - The design employed is the

 

Campbell and Stanley (1966) Pretest-Posttest Control Group
Design with two modifications: (a) a delayed posttest was

given for the purpose of reliability assessment, and (2)

19
the "experimental" and "control" groups served a double
function: that group who read the nuclear message served
as the control group for those who read the military mes-
sage; likewise, that group who read the military message
served as the control group for those who read the nuclear
message. This design lead to the collection of two data
sets to be used for the testing and comparison of the three
models of communication and belief change: one for the
nuclear message and one for the military message.

Procedures — Subjects were solicited from the Commun-

 

ication Department's Subject Research Pool, and a total of
134 subjects participated. Each subject took the pretest,
the posttest, and the delayed posttest.

Upon entrance to the experimental laboratory, the sub-
jects were given the questionnaire that appears in Appendix
D. They were asked to quietly read the instructions and to
raise their hand if clarification was needed. Upon the com-
pletion of this questionnaire, the questionnaire that
appears in Appendix E was administered. Nested within the
second questionnaire was one of the experimental messages.
Each message was randomly attached to the second question-
naire, thus randomization was accomplished. The second
questionnaire which contained one of the experimental mes-
sages also contained instructions which asked each subject
to carefully read the enclosed message (article) and to

underline the key points of the article (see Appendix D).

20
One week later, the subjects returned and took the delayed
posttest.

Measurement - Accumulated information scales similar

 

to those used in the pilot study were used in the main study;
however, for the purpose of reliability assessment multiple
indicators were employed. Four informed scales and four ex-
posure scales were used and they appear in Appendix D. The
informed scales were characterized by the following descrip-
tors: "know a little-know a lot," "not aware-aware," "not
informed-informed," and "not knowledgeable-knowledgeable."
And, the exposure scales emphasized message exposures to the
following informational sources: television and radio, news-
papers and magazines, books, and interpersonal contacts.
Score value assignment was identical to that in the pilot
study; however, a different log transformation was used:

ln (x + 1); this was done for two reasons: (a) to insure
that zero exposure would be scored as zero, and (b) to ex-
pand the potential range from 2 to 4.7.

To obtain multiple measures of belief, three descriptors
were used in two formats resulting in six belief indicators
for each of the two topics. One format was similiar to that
used in the pilot study; this consisted of assigning un-
likely, improbable, and false = 0 and likely, probable, and
true = 100. With this scale, each subject was asked to
report his or her belief by choosing the number which best

represented his or her belief.

21

The second scaling format for the measurement of be-
lief used the identical descriptors, but in this case, the
subjects were asked to make a forced choice between the two
end-points. After their selection, they were then asked to
estimate how certain they were by marking a Likert-type
scale with the following end points: "just guessing-certain."
For those who marked "just guessing" received a score of 50.
If they initially choose "true," the score value was assigned
in decrements of ten so that "certain" equaled 0. If they
initially choose "false," the score was assigned in increments
of ten so that "certain" equaled 100. Thus, for the analyses
of the belief indicators were scored so that "true" was set

to zero and "false" to 100.

CHAPTER III

PRELIMINARY ANALYSES

In this chapter, the results of the scale construction
analyses and procedures are presented. The reliability of
the resulting scale measures are then reported as internal
consistency coefficients and reliability coefficients as
derived by an over-time estimation procedure. Lastly, a
check on the effectiveness of the message manipulation is

reported.

Scale Construction, Evaluation, and Reliability

 

Since multiple indicators for the belief and accumulated
information constructs were used, this permitted each of the
indicators to be evaluated in terms of their relation to an
underlying "unmeasured" variable. To assess the quality of
each of the multiple indicators, an oblique multiple groups
cluster analysis was performed on the pretest data (Hunter &
Cohen, 1969). All of the indicators reported in Appendix D
were cluster analyzed: the six indicators for military (and
for nuclear) belief, and the eight indicators for military
(and for nuclear) accumulated information. Logrithmic trans-
formations were performed on all of the exposure indicators
before the analysis of clusters.

22

23

The cluster matrix appears in Appendix G; the results
of this analysis indicated that the belief indicators were
(1) consistently intercorrelated with themselves, and (2)
consistently correlated with indicators outside of the be-
lief cluster. The accumulated information indicators were
also consistently intercorrelated; however, they do show
deviations in their correlations with outside indicators:
the exposure and the informed indicators show slight devia-
tions in outside correlations. This suggested that these
two sets of indicators might be better treated as two
separate clusters.

To evaluate whether the accumulated information indi-
cators should be treated as two separate clusters, these
two sets of indicators were separated and a second cluster
analysis was performed. The results of this analysis yield-
ed cluster correlations between the exposure and the inform-
ed clusters to be .73 (for the military indicators) and .81
(for the nuclear indicators). As such, an accumulated
information measure using both the exposure and the informed
indicators of accumulated information was constructed. All
of the values for both the belief and the accumulated in-
formation measures reported below are based upon scales
constructed by the simple average of the indicators.

The reliability of these measures was estimated in two
ways: (1) coefficient alphas (internal consistency) coef-

ficients were computed for each of the measures from the

 

24
pretest data; (2) an over-time reliability procedure was
used using the data collected at the three points in time
(Wiley & Wiley, 1970). For this later over time procedure,
only the data obtained from the control situations were
used. The obtained reliability coefficients appear in
Table 1.

Table 1. Reliability Estimates for the Pretest Measures
of Belief and Accumulated Information

 

 

Measure Topic Reliability Estimates
Belief Nuclear .97(.90)*
Military .96(.93)
Accumulated Nuclear .94(.98)
Inf°rmat1°n Military .91(.94)

 

*

The first entry is the internal consistency coefficient
and the entry in parenthesis is the over-time reliability
estimate.

Message Effect Manipulation Check

 

The means and the standard deviations for the pretest,
posttest, and the belief changes are presented in Table 2.
Table 2. Mean Pretest, Posttest, and Change Values and

Standard Deviations for the Experimental and
Control Situations

 

 

Message Belief Belief
Treatment Topic Pretest Posttest Change
Nuclear Nuclear 33.7(21.7)* 21.1(26.4) -12.6(18.7)
Control Nuclear 35.4(27.0) 35.8(26.7) 0.4(15.3)
Military Military 38.9(26.2) 34.3(26.4) - 4.6(16.3)
Control Military 43.6(23.7) 46.0(26.4) 2.4(14.5)

 

*
The value in parenthesis is the standard deviation.

25

For those who read the nuclear message, there was a
mean change of -12.6 units; for those who did not read this
message, the change on the nuclear belief topic equaled .4
units. For those who read the military message, there was
a mean change of -4.6 units; for those who did not read
this message, the change on the military belief topic
equaled 2.4.

To evaluate whether the experimental messages had the
desired impact upon belief change, the messages were dummy
coded and used to predict the belief changes as revealed by
the pretest and posttest. This was done for both the "nu—
clear-control" and the "military-control" situations. The
resulting regression analysis (summarized in Tables 3 and
4) revealed that the messages did have the effect desired.
The correlation between the dummy message variable and be-
lief change for the nuclear message equaled .36 and the
correlation between the dummy message variable and belief

change for the military message equaled .21.1

Table 3. ANOVA Regression Summary for Effects of Nuclear
Message upon Nuclear Belief Change

 

 

Source df SS MS r F sig.
Nuclear
Message 1 5678.85 5678.85 .36 19.62 p <.001

Residual 132 38197.06 289.40

 

26

Table 4. ANOVA Regression Summary for Effects of Military
Message upon Military Belief Change

 

 

Source df SS MS r F sig.
Military
Message 1 1542.70 1542.70 .21 6.42 p=.01

Residual 132 31722.70 240.30

 

CHAPTER IV

RESULTS

Testing the Models

 

Each of the three models has the form
Ab = ~dd

where g is either the message-belief discrepancy or a mod-
ification of that discrepancy. In each case d can be cal-
culated from the other variables. Thus, one quick check

of the relative power of the three models is to compare the
correlation r for each of the three models. These cor-

dAb

relations are presented in Table 5.

Table 5. Correlations Between Ab and d for the Nuclear
and Military Topics in Both the Experimental
and Control Situations

 

hhchfiur lmxflear Nfilhwmy Nfikflxuy

 

variable Message Control .Message Cbntrol
anxmtkmal

Change d = b0 -.65 -.30 -.30 —.27
Aommmflabai

Information d = bO/i0 -.76 -.26 —.38 -.12
Belief

Certainty d = bO/BcO -.16 -.13 -.08 .07
Accumulated

Information 10 .51 .04 .32 .16
Belief

CErtainty BcO .39 .26 .14 .06

 

27

28
The first column of Table 5 has the correlations for
the nuclear message group. The correlation for the propor-

tional change model in which d = b is -.65 which is sub-

0

stantial. In part, this is the well known regression
artifact, but only in small part. The control group cor-
relation for change on the nuclear belief is found in column
two of Table 5 and is only -.30 and that represents the max-
imum value of this correlation that could be created by a
regression artifact. The correlation for the accumulated
information model in which d = (bO/io) is -.76 for the nu-
clear message group which is not only substantial in size,
but is larger than the -.65 for the proportional change
model. Thus belief change for those whose belief is based
upon more accumulated information. The correlation for the
belief certainty model in which d = (ECO/b0) is .16 which

is negligible in comparison to the fit for the other two
models.

The correlations for the military message group are
presented in column three of Table 5 and the corresponding
control group correlations are found in column four. The
correlation for the proportional change model is -.30, the
correlation for the accumulated information model is -.38,
and the correlation for the belief certainty model is -.08.
These correlations are all lower than those obtained for
the nuclear message group and reflect a difference in the

basic effectiveness of the two messages. However, the

29
comparative size of the correlations is the same: the be-
lief certainty model shows almost no fit at all, while the
accumulated information model shows definitely better fit
than the proportional change model.

However, there is one question which the correlations
in Table 5 cannot answer: Might there be an even better
functional form for the accumulated information model that
hypothesized before the data was gathered? To answer this
question a more general graphic method was used to analyze
belief change as a function of initial belief and accumu—
lated information.

For the analyses presented below, initial belief (b0)
was divided into the following three levels: 0-35 = 1,
35.001-65 = 2, and 65.001-100 = 3. And, initial accumu-
lated information was divided into the following three
levels: 0-l.75 = l, 1.75001-3.25 = 2, and those values
greater than 3.25001 were scored as 3. For the nuclear
message, the belief change means and number of subjects are
presented in Table 6.

The repression of belief change (Ab) onto the three
levels of initial belief (b0) and the three levels of
initial accumulated information (10) produced the following

parametric curves for the nuclear message (see Figure 4):

30

 

 

 

Table 6. Belief Change Means and Sample Size for the Three
Levels of Accumulated Information and the Three
Levels of Belief for the Nuclear Message
Accumulated Information (i0)
1 2 3
I I
3 ' —53.9( 2)* -26.3( 2) -ll.7(l) ' -34.4( 5)
Belief 2 ' —27.l(18) —1l.7( 5) - 2.7(1) ' -22.9(24)
(b ) I I
0 1 . - 7.0(11) - 1.0(21) - 2.1(5) . - 2.9(37)
I __________________ I ______
' -21.7(31) - 4.7(28) - 3.6(7) ' -12.6(66)
I I
*
The values in parenthesis refer to cell size, n = 66.

 

 

/\
Ab 2 3 \_b0
hxi I
-10 fi\\\\\\\\,
10:3
-20
i0=2
-30
-40
-50
10:1
-60

 

Figure 4.

Regression curves
change on initial
lated information

of nuclear belief
belief with accumu-
parameterized

31
The results of the regression analysis reported in Figure 4
clearly support the accumulated information model. The
parametric curves almost exactly reproduce the accumulated
information parametric curve reported in Figure 2 in
Chapter I.

The same analysis was performed for the military mes-
sage; the belief change means and number of subjects are
presented in Table 7. The resulting regression plot
appears in Figure 5.

Table 7. Belief Change Means and Sample Size for the Three

Levels of Accumulated Information and the Three
Levels of Belief for the Military Message

 

Accumulated Information (10)

 

1 2 3

3 I -25.0( 3) -12.7( 4) 0.3(6) I -8.3(l3)

Belief 2 I -10.8(l7) 6.0( 5) --(o) I -7.9(22)
(b ) I I

0 1 I - 6.9( 6) 0.0(22) -o.2(5) I -1.3(33)

' __________________ .L _____

I -ll.5(26) - .07(31) .007(11) I -4.6(68)

 

*
The value in parenthesis refer to cell size, n = 68.

The results of this analysis for the military message
supports the accumulated information model; however, the
support received is not as clean as that produced by the
nuclear message. One reason for the "rougher" looking para-
metric curves is due to the fact legs change occurred with
the military message. The belief and accumulated information

measures for both the nuclear and military topics were about

32

 

 

 

 

A\
10
Ab ,,//”//4\\ 4410?? b0
0—— /
Q
'10 \
'i =2
0
-20
~30 10:1
-40
Figure 5. Regression curves of military belief

change on initial belief with accumulated
information parameterized
equal in reliability; thus they shared about equal errors
of measurement. These errors tend to become increasingly
pronounced when smaller changes occur; with larger changes,

they are there, but they are less obvious to the eye.

Accumulated Information and Belief Certainty

 

The discussion in Chapter I suggested that there would
be a correlation between accumulated information and belief
certainty. With the pretest data, the correlation equaled
.56 for the military topic (.67 if corrected for attenua-
tion); this correlation equaled .62 for the nuclear topic

(.70 if corrected for attenuation).

33

However, this way of viewing the relationship between
belief and accumulated information is somewhat misleading
since it passes over the basic nonlinear relationship be-
tween accumulated information and belief. To draw out this
nonlinear relationship, the same intervales for belief and
accumulated information were used as were used in the pre-
ceding discussion of belief change (refer to discussion
preceding Tables 6 and 7). Table 8a presents the number of
persons in each of the cells for the three levels of belief
and the three levels of accumulated information for the pre-
tested nuclear beliefs and accumulated information. Table
8b presents the same, but for the military topic, and
Table 8c presents both topics combined.
Table 8a. Number of Subjects in Each of the Cells for the

Three Levels of Belief and the Three Levels of
Accumulated Information for the Nuclear Pretest

 

 

Group
Belief(b0)
l 2 3

Accumulated ' "
Information 3 ' 7 l 17 " 25

(i0) I II
2 ' 8 9 35 " 52

I II
1 ' 4 24 29 " 57

 

34

Table 8b. Number of Subjects in Each of the Cells for the
Three Levels of Belief and the Three Levels of
Accumulated Information for the Military Pretest

 

 

Group
Belief (b )
o
1 2 3
Accumulated ' "
Information 3 ' 7 0 8 " 15
(i0) ' u
2 ' 11 13 37 " 61
l ' 8 36 14 " 58
' 26 49 59 " 134

 

Table 80. Number of Subjects in Each of the Belief-
Accumulated Information Cells for Both the
Nuclear and Military Topics

 

 

, Belief(b0) u

. l 2 3 n

I ll

Accumulated ' "
Information 3 ' l4 1 25 " 40

(i0) I ll
2 ' 19 22 72 " 113
l ' 12 50 43 " 115

I

45 83 140 " 268

 

Table 8a and 8b shows the pretest relation between
accumulated information and belief certainty for the nuclear
topic and for the military topic. Since they are similar,
the distribution was pooled and presented in Table 8c. The

top row of Table 8c shows the distribution of belief among

35

persons who are highly informed. Only 1 out of 40 persons
is neutral. Looking at the middle column of Table 8c re-
veals the same fact viewed from the other perspective: of
the 83 persons who were neutral on either the nuclear or
military topics, only 1 person regarded him or herself as
well informed. The bottom row of Table 8c shows the distri-
bution of belief among those who were poorly informed on one
topic of the other. Of these 115 subjects, 60 or 52 per
cent were neutral. However, the other 48 per cent held in-
tense beliefs in one direction or the other.

Thus the data revealed only 1 person in 268 who was an
"informed neutral," but did find 55 out of 268 who were "un—

informed resolutes."

CHAPTER V

DISCUSSION

The prediction of belief change was enhanced with the
inclusion of the accumulated information operator, a finding
which supports the Saltiel-Woelfel (1975) hypothesis. In
this model, the amount of change obtained is inhibited by
the amount of information accumulated into the belief.
Higher amounts of accumulated information yield lesser
changes and lower amounts of accumulated information yield
greater changes.

Other literature also supports this finding. Patterson
and McClure (1973) in a political mass communication study
assessed the impact of paid political message upon belief
change. The message and belief evaluated was a "Democrats
for Nixon" political message which argued that: (a)

Richard Nixon does not favor spending less money on the
military, and (b) George McGovern favors spending less money
on the military. The results they obtained appear in

Table 9.

36

37

Table 9. Mean Belief Changes as a Function of Prime Time
Television Exposure (from Patterson & McClure,

 

 

1973)
Military Spending Military Spending
Exposure (McGovern) (Nixon)
High TV
Viewers 29 19
Low TV
Viewers ll 10

 

They found that potential exposure to the commercial message
(as measured by prime time TV exposure) lead to belief
changes in the direction of the message. Also, they found
that the commercial message had less of an effect on beliefs
about Nixon than on beliefs about McGovern. At the time of
the study, Nixon had been president for more than three
years, a vice president for four years, and was in the news
frequently. McGovern on the other hand was well known only
in his home state of South Dakota. Thus, it is very likely
that many receivers had already accumulated much information
about Nixon's relationship to the military, and very little
about McGovern's relationship.

Could these findings be explained by the proportional
change model? True, if more persons already believed that
"Nixon does not favor spending less money on the military,"
then there would be less distance to move and hence less
possible change. However, the way in which Patterson and

McClure (1973) calculated belief change suggests that this

38
is not so. These researchers obtained belief change values
by subtracting the per cent of persons who changed in the
direction of the message from the per cent of persons who
changedin the direction away from the message. Thus, the
amount of change was not considered, but only whether a
person changed.

The belief certainty model clearly was the most in-
ferior of the three models. It failed, and its failure
indicates that the certainty to which one ascribes to a
belief is unrelated to belief change, a finding that was
contrary to the expectations of this study. Overall, the
results of this study indicate that belief change resistance
is best viewed as being linked to an informational base,

rather than belief strength.

FOOTNOTE

Special care was taken to insure high reliability of
measurement. For each variable multiple indicators
were used: six for the belief measures and eight for
the accumulated information measures. Reliability
was also double checked for accuracy; in each case
the reliability estimates were in the .90's. Thus,
problems related to the use of change scores were
minimized and correction procedures such as those
suggested by Hunter and Cohen (1974) were not
necessary.

39

APPENDIX A

Belief topics and standard deviations for the
belief certainty and accumulated information
measures used in the pilot study.

APPENDIX A

Belief topics and standard deviations for the belief cer-
tainty and accumulated information measures used in the
pilot study.

STANDARD DEVIATIONS FOR BELIEF CERTAINTY AND THE
INFORMED AND EXPOSURE INFORMATION MEASURES (N = 33)

 

Belief In- Expo-
Belief Topic Certainty formed- sure

 

1. The price of gold is well
over $100 per ounce. 49.07 1.85 .55

2. U.S. citizens are drinking
less coffee today than in
the 1960's. 36.63 1.58 .62

3. Two out of three persons who
run away (i.e., desert or
leave without a divorce or
separation) from their hus-
bands or wives are women. 31.67 1.83 .67

4.* The U.S.S.R. military forces
are becoming superior to the
military forces of the U.S.A. 39.60 1.95 .74

5. Since 1962, the number of
persons taking the IRS (In-
ternal Revenue Service) to
court has almost tripled. 31.80 1.33 .54

6. Recently, the enrollment in
theological seminaries
(religion schools) has
increased. 34.43 1.78 .66

7. About 500,000 U.S. citizens
make crime and criminal
activities their sole
source of economic support. 39.76 1.65 .67

8. Large numbers of Jews are
currently fleeing Israel. 35.60 1.58 .69

k

*
Denotes belief topic selected for the experiment.

40

41

APPENDIX A (cont'd.)

 

Belief Topic

Belief
Certainty formed

In-

Expo-
sure

 

10.

11.

12.

13.

14.

15.

16.

17.

The U.S. is importing more
foreign oil now than ever
before.

The nuclear production of
electricity is potentially
more dangerous than the con-
ventional methods of pro-
ducing electricity.

More and more U.S. citizens
are moving to the south-
western states.

Women are committing more

crimes and are engaged in more

criminal activities than ever
before.

The economic (trade) alliance
between Communist China and
the U.S.A. is becoming
stronger.

Fidel Castro of Cuba and Prime
Minister Pierre Trudeau of
Canada are close personal
friends.

Of all the countries in the
world, within the last few
years most of the immigrants

legally admitted to the U.S.A.

were from Mexico.

The U.S. military forces are
virtually useless against
Cuba.

About 40% of Russian diplomats
and scientists visiting and
working in the United States
are actually Russian Spies.

 

*

34.31

34.99

34.13

38.05

37.11

37.57

33.66

39.50

38.70

1.69

1.94

1.45

1.11

Denotes belief topic selected for the experiment.

.69

.64

.61

.67

.50

.34

.59

.71

.53

42

APPENDIX A (cont'd.)

 

 

Belief In- Expo-
Belief Topic Certainty formed sure
18. The big cities of the U.S.
are currently losing large
numbers of people. 37.30 1.99 .61

 

One of the experimental messages that appears in
either Appendix B or in Appendix C was inserted
in this section of the posttest.

APPENDIX B

The military message used in the experiment.
The material was adapted from an actual news
story presented in the March 8, 1976 issue
of Time: "That Alarming Soviet Buildup."

APPENDIX B
The military message used in the experiment. The material
was adapted from an actual news story presented in the March
8, 1976 issue of Time: "That Alarming Soviet Buildup."
THAT ALARMING SOVIET BUILDUP

(Adapted from Time magazine, March 8, 1976)

Whenever it has come to a choice between swords and
plowshares, Moscow has seldom hesitated. Swords it has been.
Neither bad harvests, shortages or consumer goods nor the
spirit of detente with the West has braked the Soviet Mili-
tary buildup. In the past decade, Moscow has so expanded its
forces that most defense analysts around the world agree that
the U.S. is currently losing its military superiority.

9a laad, the Soviet generals command 2.5 million men,
supported by 40,000 tanks, compared with 790,000 men and
10,000 tanks in the U.S. Army.

£2,222 alg, the U.S. has 463 heavy bombers (mostly aging
B-525) to only 135 Soviet turbojet Bisions and turboprop Bears.
But the Soviet air space is the most intensively defended in
the world: 5000 radar stations, 2,600 fighter interceptors,
12,000 highly accurate antiaircraft missiles. By contrast,
U.S. air defense has been cut back by Congress. There are
only 12 squadrons of F-106 fighters--mostly assigned to Air
National Guard--with primary mission of interception Soviet

Bombers. With large scale production already underway of

43

44
Backfire--a new, supersonic Soviet intercontinental war-
plane--the U.S.S.R. will narrow the bomber gap. Meanwhile,
the U.S. Air Force's supersonic B-l is still in the develop-
ment stage and faces stiff Opposition in Congress.

AS aaa, the Soviet admirals deploy the world's largest
naval force. The Soviets enjoy clear superiority in attack
submarines (253 vs. 73 for the U.S.), cruisers and destroyers
armed with ship-to-ship missiles (40 vs. 0 for the U.S.) and
supply ships (2,358 vs. 1,009 for the U.S.). The Soviet Navy,
however, would have trouble rushing troops and plans to in-
tervene in sudden political or military crises far from the
U.S.S.R. The U.S. has more bases abroad and can act quickly
because of its 14 attack carriers (the Soviets have none).

la nuclear forces, the Soviets boast 1,603 interconti-

 

nental balistic missiles (the U.S. has 1,054) and 54 atomic
submarines armed with nuclear warheads (to 41 for the U.S.).
Nonetheless, the U.S. has a powerful strategic deterrent. A
head start in developing launchers with multiple, indepen-
dently targetable warheads (MIRVs), combined with the striking
power of B-52 bombers, gives the U.S. a 9,000 versus 3,200
advantage in deliverable nuclear bombs and warheads. However,
Soviet improvements of nuclear deliver systems could wipe out
the American advantage. Reason: the very powerful Soviet
rockets, especially the SS-9, SS—18, and 58-19, are capable
of launching more individual warheads than can the relatively

lightweight U.S. Minuteman III missiles.

45

The U.S. also has, in certain areas, an advantage
created by SOphisticated technology. American missiles are
more accurate than their Soviet counterparts, American sub-
marines are less noisy (thus more difficult to detect), and
the U.S. Air Force F-4s and F-lSs are more versatile and
powerful than the Russian MIG 213 and MIG 235.

What worries the analysts is that this superiority may
not last, since the Soviets are determined to narrow the
quality gap, for the U.S.S.R. is devoting increasing amounts
of money to military defense. Pentagon experts estimate that
last year's Soviet military budget was $141 billion; the U.S.
spent $94 billion. In 1964 the U.S. spent $110.4 billion on
defense (in 1976 dollars) while the Soviets spent $100 billion.

Western experts offer several theories for the massive
Soviet build up: (1) to counter potential trouble along the
4,500-mile border with Russia's Marxist archenemy, China: (2)
to maintain hegemony over Eastern Europe; (3) to overcome an
"inferiority complex" vis—a—vis the U.S. that was aggravated
when Moscow had to back down during the 1962 Cuban Missile
Crisis; (4) to provide additional arms for its adventurous
clients abroad (e.g., Angola).

Most ominous is the possibility that the Soviets, despite
detente, have not yet given up their long—term goal of domi—
nating all of Europe, by force if necessary. Indeed, Soviet
forces are no longer primarily defensive as they were in the

mid-1960's. Much of the modern Russian weaponry-—from

46

missiles to tanks to fighter bombers--is offensive, aimed at
a blitzkrieg attack with quick victory as its goal. In
Central Europe, the Soviets have concentrated huge numbers
of fast tanks and powerful artillery; at sea, the Soviet
Fleet's ship-to-ship missiles could deal fatal, surprise
blows to Western warships.

Whether the Soviets actually plan to attack the Western
world, one thing is clear according to NATO Commander-in-
Chief General Alexander Haig, "the massive Soviet build up
clearly indicates that the U.S.A. is becoming the weaker of

the two military giants."

APPENDIX C

The nuclear message used in the experiment. The
material was adapted from an actual news story
presented in the March 8, 1976 issue of Time:
"The Struggle Over Nuclear Power."

APPENDIX C
The nuclear message used in the experiment. The material
was adapted from an actual news story presented in the March
8, 1976 issue of Time: "The Struggle Over Nuclear Power."

THE STRUGGLE OVER NUCLEAR POWER

(Adapted from Time magazine, March 8, 1976)

Around the nation, power company officials at press con-
ferences and on podiums have been presenting figures to show
that the nuclear production of electricity is more practical
than the conventional means of producing electricity. In
private they concede, however, that nuclear power is poten-
tially more dangerous than the conventional sources of power.

In Washington last week, a parade of executives, engineers,
and federal officials trOOped before a joint congressional
committee to rebut the charges that their installations are
unsafe and to convince an increasingly anxious American public
that nuclear plants are necessary.

The major reason for their concern is due to the increased
momentum gathered by the anti-nuclear forces. Last month a
trio of middle-level engineers at GE's nuclear energy division
in San Jose, California, suddenly resigned their jobs in pro—
test. The trio, Dale Bridenbrauh, 44, Gregory Minor, 38, and
Richard Hubbard, 38, announced that they would instead work
full time for Project Survival, the organization coordinating

an anti-nuclear drive in California. Another nuclear engineer,

47

48

Robert Pollard, 36, quit his job with Federal Nuclear Regu-
latory Commission in protest over conditions at Consolidated
Edison's Indian Point nuclear power plants in Buchannan, New
York.

All four engineers cited the same basic reasons for their
resignations: The nuclear production of electrical power is
unsafe and hazardous. The public, said the three engineers
from San Jose in a statement to the Congressional Joint Com-
mittee on Atomic Energy, "has the right to know that an
electrical appliance, such as a toaster or hair dryer, has
more stringent safety checks than the electrical instruments
that control a nuclear plant."

Threat a: Meltdown. What concerns these nuclear engineers
--and many of their fellow protesters--is not any possibility
that a nuclear plant will blow up in a mushroom cloud and
wipe out a city. What is feared, however, is a "meltdown,"
which can occur if a reactor loses the water used to control
the temperature of its uranium core. The four nuclear
engineers claim that safety systems designed to prevent ac-
cidents have not undergone enough testing. If they failed in
a crisis, say the four, the result could be disasterous.

They could indeed. An uncontrolled core would build up
heat, melt and drOp to the bottom of its container. Its heat
would vaporize whatever water remained and the pressure of
the resulting steam could burst the containment vessel and

rupture the outer reaction container as well. This could

49

release a radioactive cloud that would drift where ever the
wind blew it. Depending upon the location of the plant,
such an accident could result in numerous immediate deaths
from radiation and even more later, caused by radiation-
induced cancers.

The possibility of a "meltdown" occurring is great.
Last March, the worlds largest nuclear power plant, located
at Brown's Ferry, Alabama, was well into the chain of events
that could lead to a meltdown after human error caused fail-
ure of several key safety systems. On a lesser level, a
Northeast Utilities plant in Waterford, Connecticut spilled
radiation outside the plant when a steam condenser ruptured.
Other nuclear power plants have had to suspend operations for
anywhere from several weeks to several months as a result of
equipment failure.

The nuclear production of electrical power also raise
other questions and fears. One is that the expansion of the
nuclear power industry would make it easy for terrorists to
steal fissionable materials for homemade bombs. Another con-
cern is nuclear wastes, one of which, plutonium, has a half-
life of over 24,000 years. Safeguarding wastes alone says
Biologist Barry Commoner, would require the creation of a
kind of permanent "nuclear priesthood," to watch over the
radioactive legacy each generation of Americans handed down

to its successors.

50
To those in the anti-nuclear camp, the danger is
obvious the conclusion is clear, "The nuclear production
of electrical power poses a severe threat to the lives and

health of millions of Americans," says nuclear engineer

Richard Hubbard.

APPENDIX D

The pretest questionnaire. Administered on
May 13, 1976.

APPENDIX D

The pretest questionnaire. Administered on May 13, 1976.

DEPARTMENT OF COMMUNICATION

Participation Credit Record

Students: Please complete information below to insure you
receive credit for participating in this study.

Name

 

Student Number

 

Class to which credit is to be applied

 

Section number

 

Instructor

 

Researcher: Please supply below information identifying
study and amount of participation to be
credited to the student. Return completed
forms to Bruce Schreiman.

51

 

‘M._.______..—

52

MICHIGAN STATE UNIVERSITY East Lansing, Michigan 48824

 

College of Communication Arts Cable: Commdept
Department of Communication

The Department of Communication is currently interested in
your beliefs, your communication experiences, and your
knowledge that you may or may not have about a variety of
issues and tOpics.

 

 

We appreciate the fact that you are responding to this study;
aa’ ou cafiT ThefE are no right or wrong answers to these
questions. Your responses will in no way adversely affect
your standing in your Communication course. Further, all of
your responses are strictly confidential.

 

 

 

While answering the questions that follow, PLEASE
DO NOT TALK WITH YOUR NEIGHBOR. ANSWER THE QUESTIONS
INDEPENDENTLY, AND ANSWER THEM AS HONESTLY AS YOU CAN.

l£_you have any questions about the scales, questions
g£,t0pics, please raise your hand.

 

 

 

 

(l)

(2)

(3)

(4)

(5)

(6)

53

 

The U.S.S.R. MILITARY FORCES are becoming SUPERIOR
to the MILITARY FORCES of the U.S.A.

 

On the scale below, what number best represents your
belief?

Totally Unlikely Uncertain Totally Likely
0...10...20...30...40...50...60...70...80...90...100
From 0-100, write any number you wish

If you had to choose, would you guess the above state-
ment to be:

LIKELY or UNLIKELY (please circle one)

How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

 

The U.S.S.R. MILITARY FORCES are becoming SUPERIOR
to the MILITARY FORCES of the U.S.A.

 

On the scale below, what number best represents your
belief?

Totally improbable Uncertain Totally probable
O...10...20...30...40...50...60...70...80...90...100
From 0-100, write any number you wish

If you had to choose, would you guess the above state-
ment to be:

PROBABLE or IMPROBABLE (please circle one)

How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

 

 

~_1 .,___-

 

 

54

 

The U.S.S.R. MILITARY FORCES are becoming SUPERIOR
to the MILITARY FORCES of the U.S.A.

 

(7) On the scale below, what number best represents your
belief?

Totally False Uncertain Totally True
0...10...20...30...40...50...60...70...80...90...100
From 0-100, write any number you wish

(8) If you had to choose, would you guess the above state-
ment to be:

TRUE or FALSE (please circle one)

(9) How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

On the tOpic of:

 

The SUPERIORITY or INFERIORITY of the U.S.S.R. MILITARY
FORCES relative to the MILITARY FORCES of the U.S.A.

 

 

 

 

(10) I know a: : : : : : :I know a lot

little
(11) I am not very: : : . : : :I am very aware

aware of the of the issue
issue

(12) I am not very: : : : : : :I am very in-

informed formed about

about the the issue

issue

(13) I am not very: : : : : :I am very knowl-
knowledgeable edgeable

 

55

On the topic of:

 

The SUPERIORITY or INFERIORITY of the U.S.S.R. MILITARY
FORCES relative to the MILITARY FORCES of the U.S.A.

 

(14) How MANY TIMES have you HEARD about this t0pic on
TELEVISION or RADIO?

 

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_ eight(8) nine(9)
ten(lO) 11-25 26-50 51-75 76-100

more than 100

(15) How MANY ARTICLES have you READ about this topic in
NEWSPAPERS or MAGAZINES?

 

 

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)__ eight(8) nine(9)
ten(lO) ll-ZS 26-50 51-75 76-100

more than 100

(16) How MANY BOOKS have you READ on this topic?

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_. eight(8) nine(9)
ten(10) ll-25 26-50 51-75 76-100

more than 100

(17) How MANY TIMES have you HEARD about this topic from
INTERPERSONAL SOURCES?

 

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)__ eight(8) nine(9)
ten(10) ll-25 26-50 51-75 76-100

more than 100

(l)

(2)

(3)

(4)

(5)

(6)

56

 

The NUCLEAR production of ELECTRICITY is potentially
more DANGEROUS than the CONVENTIONAL methods of pro-
ducing ELECTRICITY

 

On the scale below, what number best represents your
belief?

Totally Unlikely Uncertain Totally Likely
0...lO...20...30...40...50...60...70...80...90...lOO
From 0-100, write any number you wish

If you had to choose, would you guess the above state-
ment to be:

LIKELY or UNLIKELY (please circle one)

How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

 

The NUCLEAR production of ELECTRICITY is potentially
more DANGEROUS than the CONVENTIONAL methods of pro-
ducing ELECTRICITY

 

On the scale below, what number best represents your
belief?

Totally improbable Uncertain Totally probable
0...lO...20...30...40...50...60...70...80...90...100
From 0-100, write any number you wish

If you had to choose, would you guess the above state-
ment to be:

PROBABLE or IMPROBABLE (please circle one)

How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

(7)

(8)

(9)

(10)

(ll)

(12)

(13)

57

 

The NUCLEAR production of ELECTRICITY is potentially
more DANGEROUS than the CONVENTIONAL methods of pro-
ducing ELECTRICITY

 

On the scale below, what number best represents your
belief?

Totally False Uncertain Totally True
0...10...20...30...40...50...60...70...80...90...100
From 0-100, write any number you wish

If you had to choose, would you guess the above state-
ment to be:

TRUE or FALSE (please circle one)

How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

On the tOpic of:

 

The SAFETY or DANGER of the NUCLEAR production of
ELECTRICITY relative to the CONVENTIONAL methods
of producing ELECTRICITY

 

I know a: :I know a lot

little

.0
.0
O.

 

I am not very: . :I am very aware
aware of the of the issue
issue

 

I am not very:

. :I am very in-

 

informed formed about
about the the issue
issue
I am not very: : : : : : :I am very knowl-

 

knowledgeable edgeable

58

On the topic of:

The SAFETY or DANGER of the NUCLEAR production of
ELECTRICITY relative to the CONVENTIONAL methods
of producing ELECTRICITY

 

(14) How MANY TIMES have you HEARD about this topic on
TELEVISION or RADIO?

 

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_ eight(8) nine(9)
ten(lO) 11-25 26-50 51-75 76-100

more than 100

(15)- How MANY ARTICLES have you READ about this topic in
NEWSPAPERS or MAGAZINES?

 

 

none(0) one(l) two(2) three(3) four(4)
five(5) six(6) seven(7)_ eight(8) nine(9)
ten(10) ll-25 26-50 51-75 76-100

more than 100

(16) How MANY BOOKS have you READ on this topic?

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_ eight(8) nine(9)
ten(lO) ll-ZS 26-50 51-75 76-100

more than 100

(17) How MANY TIMES have you HEARD about this topic from
INTERPERSONAL SOURCES?

 

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_. eight(8) nine(9)
ten(10) 11-25 26-50 51-75 76-100

more than 100

H

59

(47) What is your sex? FEMALE or MALE (please circle one)

(48) What is your class standing? (please check one)

Freshman

 

Sophomore

 

Junior

 

Senior

 

Graduate

 

(49) What is your grade point average (GPA)? GPA

(50) How old are you? years old

APPENDIX E

Posttest questionnaire. Administered May 13, 1976.

APPENDIX E

Posttest questionnaire. Administered May 13, 1976.

DEPARTMENT OF COMMUNCIATION

Participation Credit Record

Students: Please complete information below to insure you
receive credit for participating in this study.

Name

 

Student Number

 

Class to which credit is to be applied

 

Section number

 

Instructor

 

Researcher: Please supply below information identifying
study and amount of participation to be
credited to the student. Return completed
forms to Bruce Schreiman.

60

 

61

INSTRUCTIONS

This booklet contains an article condensed from Time
magazine. Please read the article carefully and underline
(with your pencil or pen) the main points of the article.

 

 

 

After you have read and underlined the main points of the
article, please fill out the attached questionnaire and be
sure to answer all of the questions contained in that
questionnaire.

 

 

 

 

 

Take as much time as you need; be sure to answer the ques-
tions as honestly as you can. After you are finished you
may leave.

 

 

REMEMBER, WE MEET AGAIN NEXT THURSDAY--SAME TIME, SAME
PLACE.

 

62

 

The U.S.S.R. MILITARY FORCES are becoming SUPERIOR
to the MILITARY FORCES of the U.S.A.

 

(1) On the scale below, what number best represents your
belief?

Totally Unlikely Uncertain Totally Likely
O...10...20...30...40...50...60...70...80...90...100
From 0-100, write any number you wish

(2) If you had to choose, would you guess the above state-
ment to be:

LIKELY or UNLIKELY (please circle one)

(3) How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

 

The U.S.S.R. MILITARY FORCES are becoming SUPERIOR
to the MILITARY FORCES of the U.S.A.

 

(4) On the scale below, what number best represents your
belief?

Totally improbable Uncertain Totally probable

0...lO...20...30...40...50...60...70...80...90...100
From O-lOO, write any number you wish

(5) If you had to choose, would you guess the above state-
ment to be:

PROBABLE or IMPROBABLE (please circle one)

(6) How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

M—

 

 

63

 

The U.S.S.R. MILITARY FORCES are becoming SUPERIOR
to the MILITARY FORCES of the U.S.A.

 

(7) On the scale below, what number best represents your
belief?

Totally False Uncertain TotallyTrue
00.0100.0200.0300004000050000600.07000080.009000.100
From 0-100, write any number you wish

(8) If you had to choose, would you guess the above state-
ment to be:

TRUE or FALSE (please circle one)

(9) How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

 

On the topic of:

 

The SUPERIORITY or INFERIORITY of the U.S.S.R. MILITARY
FORCES relative to the MILITARY FORCES of the U.S.A.

 

(10) I know a: :
little

:I know a lot

 

 

(11) I am not very: : : :I am very aware
aware of the of the issue
issue

 

 

(12) I am not very: : : :I am very in-

 

 

informed formed about
about the the issue
issue

(13) I am not very: : : :I am very knowl-
knowledgeable edgeable

 

 

64

On the topic of:

 

The SUPERIORITY or INFERIORITY of the U.S.S.R. MILITARY
FORCES relative to the MILITARY FORCES of the U.S.A.

 

(14) How MANY TIMES have you HEARD about this t0pic on
TELEVISION or RADIO?

 

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_ eight(8) nine(9)
ten(lO) ll-25 26—50 51-75 76-100

more than 100

(15) How MANY ARTICLES have you READ about this topic in
NEWSPAPERS or MAGAZINES?

 

 

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_ eight(8) nine(9)
ten(lO) ll-25 26-50 51—75 76-100

more than 100

(16) How MANY BOOKS have you READ on this topic?

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_ eight(8) nine(9)
ten(lO) ll—25 26-50 51-75 76-100

more than 100

(17) How MANY TIMES have you HEARD about this tOpic from
INTERPERSONAL SOURCES?

 

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_ eight(8) nine(9)
ten(lO) 11—25 26-50 51-75 76-100

more than 100

 

(l)

(2)

(3)

(4)

(5)

(6)

65

 

The NUCLEAR production of ELECTRICITY is potentially
more DANGEROUS than the CONVENTIONAL methods of pro-
ducing ELECTRICITY

 

On the scale below, what number best represents your
belief?

Totally Unlikely Uncertain Totally Likely
0...10...20...30...40...50...60...70...80...90...100
From 0-100, write any number you wish

If you had to choose, would you guess the above state-
ment to be:

LIKELY or UNLIKELY (please circle one)

How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

 

The NUCLEAR production of ELECTRICITY is potentially
more DANGEROUS than the CONVENTIONAL methods of pro-
ducing ELECTRICITY

 

On the scale below, what number best represents your
belief?

Totally improbable Uncertain Totally probable
0...lO...20...30...40...50...60...70...80...90...100
From O-lOO, write any number you wish

If you had to choose, would you guess the above state-
ment to be:

PROBABLE or IMPROBABLE (please circle one)

How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

(7)

(8)

(9)

(10)

(ll)

(12)

(13)

66

 

The NUCLEAR production of ELECTRICITY is potentially
more DANGEROUS than the CONVENTIONAL methods of pro-
ducing ELECTRICITY

 

On the scale below, what number best represents your
belief?

Totally False Uncertain Totally True
0...10...20...30...40...SO...60...70...80...90...100
From 0-100, write any number you wish

If you had to choose, would you guess the above state-
ment to be:

TRUE or FALSE (please circle one)

How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

On the topic of:

 

The SAFETY or DANGER of the NUCLEAR production of
ELECTRICITY relative to the CONVENTIONAL methods
of producing ELECTRICITY

 

 

 

 

I know a: : : : . . :I know a lot
little
I am not very: : : : : : :I am very aware
aware of the of the issue
issue
I am not very: : : : . : :I am very in-
informed formed about
about the the issue
issue
I am not very: : : : . :I am very knowl-

 

knowledgeable edgeable

67

On the topic of:

 

The SAFETY or DANGER of the NUCLEAR production of
ELECTRICITY relative to the CONVENTIONAL methods
of producing ELECTRICITY

 

(14) How MANY TIMES have you HEARD about this topic on
TELEVISION or RADIO?

 

none(0) one(l) two(2) three(3) four(4)
five(5) six(6) seven(7)_ eight(8) nine(9)
ten(lO) ll-25 26-50 51-75 76-100

more than 100

(15) How MANY ARTICLES have you READ about this topic in
NEWSPAPERS or MAGAZINES?

 

 

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_ eight(8) nine(9)
ten(lO) 11-25 26-50 51—75 76-100

more than 100

(16) How MANY BOOKS have you READ on this tOpic?

none(0) one(l) two(2) three(3) four(4)
five(5) six(6) seven(7)_ eight(8) nine(9)
ten(lO) 11-25 26-50 51—75 76-100

more than 100

(17) How MANY TIMES have you HEARD about this topic from
INTERPERSONAL SOURCES?

 

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_ eight(8) nine(9)
ten(lO) 11-25 26—50 51-75 76-100

more than 100

 

APPENDIX F

Delayed posttest. Administered May 20, 1976.

 

APPENDIX F

Delayed posttest. Administered May 20, 1976.

DEPARTMENT OF COMMUNICATION
Participation Credit Record

Students: Please complete information below to insure you
receive credit for participating in this study.

Name

 

Student Number

 

Class to which credit is to be applied

 

Section number

 

Instructor

 

Researcher: Please supply below information identifying
study and amount of participation to be
credited to the student. Return completed
forms to Bruce Schreiman.

68

 

69

MICHIGAN STATE UNIVERSITY East Lansing, Michigan 48824

 

College of Communication Arts Cable: Commdept
Department of Communication

The Department of Communication is currently interested in
your beliefs, your communication experiences, and your
knowledge that you may or may not have about a variety of
issues and topics.

 

 

We appreciate the fact that you are responding to this study:
all that we_a§k_i§_that you answer the questions a§_honestly
as on can. There are no right or wrong answers to these
qfiestiofigf Your responses will in no way adversely affect
your standing in your Communication course. Further, all of
your re5ponses are strictly confidential.

 

 

While answering the questions that follow, PLEASE
DO NOT TALK WITH YOUR NEIGHBOR. ANSWER THE QUESTIONS
INDEPENDENTLY,.AND ANSWER THEM AS HONESTLY AS YOU CAN.

If you have any questions about the scales, questions
95 topics, please raise your hand.

 

 

 

 

(l)

(2)

(3)

(4)

(5)

(6)

70

The U.S.S.R. MILITARY FORCES are becoming SUPERIOR
to the MILITARY FORCES of the U.S.A.

On the scale below, what number best represents your
belief?

Totally Unlikely Uncertain Totally Likely
0...lO...20...30...40...50...60...70...80...90...lOO
From 0—100, write any number you wish

If you had to choose, would you guess the above state-
ment to be:

LIKELY or UNLIKELY (please circle one)

How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

\\\\

The U.S.S.R. MILITARY FORCES are becoming SUPERIOR
to the MILITARY FORCES of the U.S.A.
.___._._______

.._______.___1_______

On the scale below, what number best represents your
belief?

Totally improbable Uncertain Totally probable
O...lO...20...30...40...50...60...70...80...90...lOO
From O—lOO, write any number you wish

If you had to choose, would you guess the above state-
ment to be:

PROBABLE or IMPROBABLE (please circle one)

How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

\\\\

 

 

(7)

(8)

(9)

(10)

(ll)

(12)

(13)

71

 

The U.S.S.R. MILITARY FORCES are becoming SUPERIOR
to the MILITARY FORCES of the U.S.A.

 

On the scale below, what number best represents your
belief?

Totally False Uncertain Totally True
0...lO...20...30...40...50...60...70...80...90...100

From O-lOO, write any number you wish

If you had to choose, would you guess the above state-
ment to be:

TRUE or FALSE (please circle one)

How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

On the topic of:

 

The SUPERIORITY or INFERIORITY of the U.S.S.R. MILITARY
FORCES relative to the MILITARY FORCES of the U.S.A.

 

I know a: : :I know a lot

little

 

O.
O.

I am not very: . . :I am very aware
aware of the of the issue
issue

 

I am not very: : : . : :I am very in-
informed formed about

 

about the the issue
issue
I am not very: : : : : : :I am very knowl-

 

knowledgeable edgeable

72

On the topic of:

 

The SUPERIORITY or INFERIORITY of the U.S.S.R. MILITARY
FORCES relative to the MILITARY FORCES of the U.S.A.

 

(14) How MANY TIMES have you HEARD about this topic on
TELEVISION or RADIO?

 

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)__ eight(8) nine(9)
ten(lO) 11-25 26-50 51-75 76—100

more than 100

(15) How MANY ARTICLES have you READ about this topic in
NEWSPAPERS or MAGAZINES?

 

 

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_ eight(8) nine(9)
ten(lO) 11425 26—50 51-75 76-100

more than 100

(16) How MANY BOOKS have you READ on this topic?

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)__ eight(8) nine(9)
ten(lO) 11-25 26-50 51-75 76-100

more than 100

(17) How MANY TIMES have you HEARD about this topic from
INTERPERSONAL SOURCES?

 

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_ eight(8) nine(9)
ten(lO) ll-ZS 26-50 51-75 76-100

more than 100

(l)

(2)

(3)

(4)

(5)

(6)

73

 

The NUCLEAR production of ELECTRICITY is potentially
more DANGEROUS than the CONVENTIONAL methods of pro-
ducing ELECTRICITY

 

On the scale below, what number best represents your
belief?

Totally Unlikely Uncertain Totally Likely
O...10...20...30...40...50...60...70...80...90...100

From 0-100, write any number you wish

If you had to choose, would you guess the above state-
ment to be:

LIKELY or UNLIKELY (please circle one)

How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

 

The NUCLEAR production of ELECTRICITY is potentially
more DANGEROUS than the CONVENTIONAL methods of pro-
ducing ELECTRICITY

 

On the scale below, what number best represents your
belief?

Totally improbable Uncertain Totally probable
0...lO...20...30...40...50...60...70...80...90...100
From 0-100, write any number you wish

If you had to choose, would you guess the above state—
ment to be:

PROBABLE or IMPROBABLE (please circle one)

How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

(7)

(8)

(9)

(10)

(ll)

(12)

(13)

74

 

The NUCLEAR production of ELECTRICITY is potentially
more DANGEROUS than the CONVENTIONAL methods of pro-
ducing ELECTRICITY

 

On the scale below, what number best represents your
belief?

Totally False Uncertain Totally True
0...10...20...30...40...50...60...70...80...90...100
From 0-100, write any number you wish

If you had to choose, would you guess the above state-
ment to be: -

TRUE or FALSE (please circle one)

How confident do you feel about that choice?

Just Guessing: : : : : : :Certain

 

One the topic of:

 

The SAFETY or DANGER of the NUCLEAR production of
ELECTRICITY relative to the CONVENTIONAL methods
of produicing ELECTRICITY

 

 

 

 

I know a: : : : : : :I know a lot
little
I am not very: : : . : : :I am very aware
aware of the of the issue
issue
I am not very: : : : . . :I am very in-
informed formed about
about the the issue
issue
I am not very: : : : : :I am very knowl-

 

knowledgeable edgeable

75

On the topic of:

 

The SAFETY or DANGER of the NUCLEAR production of
ELECTRICITY relative to the CONVENTIONAL methods
of producing ELECTRICITY

 

(14) How MANY TIMES have you HEARD about this topic on
TELEVISION or RADIO?

 

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_. eight(8) nine(9)
ten(lO) ll-25 26-50 51-75 76-100

more than 100

(15) How MANY ARTICLES have you READ about this tOpic in
NEWSPAPERS or MAGAZINES?

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_. eight(8) nine(9)
ten(lO) ll-25 26-50 51-75 76-100

more than 100

(16) How MANY BOOKS have you READ on this topic?

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_ eight(8) nine(9)
ten(lO) ll-25 26-50 51-75 76-100

more than 100

(17) How MANY TIMES have you HEARD about this topic from
INTERPERSONAL SOURCES?

 

none(0) one(1) two(2) three(3) four(4)
five(5) six(6) seven(7)_' eight(8) nine(9)
ten(lO) ll-ZS 26-50 51—75 76-100

more than 100

|

APPENDIX G

Multiple groups, oblique cluster analysis
(key to indicators)

APPENDIX G
Multiple Groups, Oblique Cluster Analysis

(Key to Indicators)

Belief (Military)

 

Unlikely-Likely (0-100)

Unlikely-Likely (Forced Choice-Interval)
Improbable-Probable (0-100)
Improbable-Probable (Forced Choice-Interval)
False-True (0-100)

False-True (Forced Choice-Interval)

mU'I-waH

Belief (Nuclear)

 

7. Unlikely-Likely (0-100)

8. Unlikely-Likely (Forced Choice-Interval)

9. Improbable-Probable (0-100)
10. Improbable-Probable (Forced Choice-Interval)
11. False-True (0—100)
12. False-True (Forced Choice-Interval)

Accumulated Information (Military)

 

13. Know a little-Know a lot

14. Not aware—Aware

15. Not Informed-Informed

16. Not Knowledgeable-Knowledgeable
1?. Exposure to Television/Radio

18. Exposure to Newspapers/Magazines
l9. Exposure to Books

20. Exposure to Interpersonal Sources

Accumulated Information (Nuclear)

 

21. Know a little-Know a lot

22. Not Aware-Aware

23. Not Informed—Informed

24. Not Knowledgeable-Knowledgeable
25. Exposure to Television/Radio

26. Exposure to Newspapers/Magazines
27. Exposure to Books

28. Exposure to Interpersonal Sources

76

.
III-II-...I ..|,-; -- .---. I - .I L -. .II
.vc .-Hu.. .mo .co
ax._1r< 52L.” LZwr mzuuu nr<=r<hm
,,- _ A «Han-H cxmzzi-,a - armou- .;stm::,vmmnII
.Hea.rH x:c:. z_.H»<1«ruz »: H H»: m w
cHnHI uLHJ..a 1=mguzr Now
-;-I.--; IIIrIIII- -;-:-I--:II-IIIIIIIIIIIII:--III- -.+IH ml .EIHIHHHH:.How-I.

 

 

‘Il...l-I|l I '-

 

 

 

 

z<zucm1 maroz: u_&~. 43:

IIIIIIII'IIIIII'III‘II I. 'I'IIII.

.ruu:.1xcc.acor

 

I
I
l
i
l
i

 

 

 

 

 

 

 

 

 

 

 

      

 

 

 

 

 

 

 

HOH :WH .H. o-» on. co :0 Ha NH rm cm ck mn km Hv c. m» ox mm Mn c- CH- k- «H- v. «H- H. a- o. H- .H- c- .cm
Hm . H H m o k v -..m e Ho Ho H» .n on v xH me n - m c- n e. HH- «u. r - .m- 1H. Hm. mmm
-RH. H -LoH-mmm .. j. c ...m.- w. mm.a «H. pm. n-I:».I-H.- a -.vw-:nm: «I -<...maIIwo-:voIInoIsmoIIo:-IHHIIr I.amI .I.-.m vi -..ncm- - I-
c- H - .H ooH a. cH. HH- HH. w o. M «- mm. cw. «n- o~n k. H» m. a. cw o oH a «H “H no no :1 km «0 Wm Hem
as 1. . . 4‘ H c - I-wIIIsIIna H cu - H. m. HH. cm- - H- - :N- c. H- HH. HH. Hm
4.- a. -t«.-mm. m» ca am 7.” “m me mu- no -am om we n~--c~ «V «a HH-.H--HH- mm. ».:.mx-muI-om-.o..lx- .H- “H-. -HHI-- .....
mm Av pH. HH. we on «H up an on no mo rn an cv so mm ex we «v a. CH- ~H- mH- A. Hm- H- mH- n. .H- cH- VH. cm
. .n HH. «H. a «0 m» «a Hm mv .0 Ho H n H . a an «N av «H . cH- ..- o. :x- - pH HH- .H- HH- “H- mm
- an: “H.7-H. a cm co mm kn . m .wc no co Hm Hm cm .nw me .nH w» a. 0-..-Mwm ;-Hm- :u.--aH- WA--c---n H. H-- H . Hm --
.m AH HH o. q. me on «v m. em mm c. o o HH «H HH em - HH “H HH .H mH mH .. .H- m. «H- . HH- H- nu
x» .n y H mo nH no no mo mm He nu «n Hn mm «a v» no He cc u n. H H. v e. H. c m. w .- H Hm
-:-dk- iv .HH.-a- chImo no no -mm cc wk am OH- «x c» mm m» an Nm no mu--eHu «HI-«Hm.nut-HHn-HHm-au-IHH..m-:-_H.-m- -- HR ,
m Hm .. a n n H , o c cH 1 o -II-c . - - - cu - - .w- . . no. H
Ha Ha "- wmo hm mm cm nw rm 0 Hm v~ a» mm mm x” «w n" mm NW N. «I -u w- m mm mm. $mn c.- no- “M- wo- c
- .v Ho, H-I-wn..kn .on-,cv..vq. -oH -mm.-mn-.cn es. en Imo. mm ;ow Ian..mm-.mv H. -n-.I«.-~- :I--.-H--.Hm-mn- aw..on. wm- mu. HH IIII-I
.v H» i em. on v. kn ov mH a. an am He Ho HH cu m. av Hm mH c- m. H H v .- ¢H- an- «H- Ho- TH. mn- kH
an on ~ s. cm n~ um co we HH en mm ev cc on me «a no em HA 9 a- HH H k H- v- a- 4. CH. H HH- em
I mm HM--IHI-H.I I-i .c-;¢~I mmII-HIIamIIwnIIm~ .mI-NvgummI,me-mw.:m- -omI-mw .H--CHIIMHIIHHII-HI-HI.-m-.IIH IIH.IIwI--HIIIH.-- m..II.I-I
- m . on . . . H .- . - H n- H
.m mo 1. m- m" m" m" ww mm «W Hm om m MH mm mm mm ”w om m“ H mH- M .H-.m- HH-H1- m- HH- “H- .- H. m.
II-.- H.s-«o-zo~a-nHu nH;- . -aHu-v.-~H Iv;;I¢. -.- q- n. IdrIIo HH,IH HI «a -Hc..H«-Ion-.waI-cH .HHIIHHIIHHI.+H-:HN- wH ...... wHI :II
.H. m. .o o ow. mm. eH. cw” ”Ha HH H. v“. c- c- H. m- - eH n- (H. an we so «a my Ha HH m H v .H H HH
nm E yo 0H 0-. HH. «H. oHH c . cH nu o c- H- w. H HH «H o n ma we we -Hm am Hm Ha mH mH a pm mH oH
a . .- IHoI - oH Hm- .H- ow «H .m. IH e:eH--H.- H. :-. H- H a -H - HH- . no .HH 3 --.-:-mx - ; --- - HI-.H H .I :o.-III-
«- m o wH A. o. M. 0. on TH w an n N m v N am A v. rm «5 «m ww w“ vi m mH m 1 HM .H r
.H- 0. Ho HH mm. «~. HN- «w. 0H. NH .. OH. H- c- HH- «- n. c H- HH- pk HH rs He He -wm HH HH HH H OH wmr H
i H. -VH. «H. mo n- mo-Isr .-w. H.--nHu-Hu.-cH--cHI;nH-.q~- eHo v. n .,H-. n. «I .H .HN HTIIHH H :1 mo re «H .o Hm o-:---
o. vm. .H no 9. HH. «H. rHu n. m. c cu sm- He- m«- «a. o. H m. «- vH c mH c nH HH rm cm HH vm on we m
c- mH. “H mo 0. o. 5. HH. ~ «H» n. Hm- oH- oH- om- ¢H- v. H» a. HH- mH m mH H m cH s: eh HR um mm «H .
-..-. .-- ;;-mm. 0. OH. H--l H:--m--m---H I-AN- NH.HH~.. 0H- ---:H-.- H-..H n:IIm,--I-:mw. vI. mmIIHm:I.s :.H “H..ww .m.--I.--
.H. «H- .w “Hu-o HH. cH. v e. VH. Ho cH. m 0H. mm- mH- N m H H. mm .mH px .H AH «H yo am we «n Hm :2 w
0. RN. HH No «H. NH. VH. ch v .n. n m. mm. mm- mm. non HH- m. m- c. cH n m_H o vH <H Hr no wk ca Hm w» H
.sm mom “om Han o~ «u cu ww HN HN km Hm c~ oH cH AH 0H mH vH HH NH HH oH c m s c,I m H r n H
II. --I-I||I-!iIEI-lIfIIIIII- .I III-.II : . Jml3c<_n.ézthW!Infi- 423r200::

meHo; vzlnccb 02¢ u ;H -- xunzwth «epocu

REFERENCES

   

 

REFERENCES

Anderson, N. Linear models for response measured on a
continuous scale. Journal of Mathematical Psychology,
1964, 1, 121-142.

 

Anderson, N. Primacy effects in personality impression
formation. Journal of Personality and Social Psy-
ch010gy, 1965, g, 1-9.

 

Anderson, N. Information integration and attitude change.
Psychological Review, 1971, 22! 171-206.

 

Anderson, N., and Hovland, C. The representation of order
effects in communication research. In C. Hovland
(ed.), The Order of Presentation in Persuasion.

New Haven: Yale University Press, 1957.

 

Bem, D. Beliefs, attitudes, and human affairs. Belmont:
Cole Publishing Co., 1970.

 

Boulding, K. The image: Knowledge in life and society.
Ann Arbor: University of Michigan Press, 1956.

 

Brim, O. Attitude content-intensity. American Sociological

 

Review, 1955, g, 68-76.

Campbell, D., and Stanley, J. Experimental and quasi-
experimental designs for research. Chicago: Rand
McNally, 1966.

 

 

Cronbach, L. Coefficient alpha and the internal structure
of tests. Psychometrika, 1951, 16, 297-334.

 

Fishbein, M., and Ajzen, I. Belief, attitude, intention,
and behavior. Reading, Mass.: AddiSon-Wesley, 1975.

 

 

French, J. A formal theory of social power. Psychological

 

Hunter, J., and Cohen, S. Mathematical models of attitude
change in the passive communICatIOn context. Un-
published’book, Michigan State University, 1972.

 

 

78

79

Hunter, J., and Cohen, S. Correcting for unreliability in
nonlinear models of attitude change. Psychometrika,
1974, 39, 445-468.

 

Hunter, J., and Cohen, S. PACKAGE: A system of computer
routines for the analysis of correlational data.
Educational and Psychological Measurement, 1969, 22,
697-700.

 

Hovland, C. I. Reconciling conflicting results derived
from experimental and survey studies of attitude
change. American PsychOlogist, 1959, lg, 8-17.

 

Lin, N. The study of human communication. Indianapolis:
Bobbs-Merrill, Co., 1973.

 

Osgood, C., and Tannenbaum, P. The principle of congruity
in the prediction of attitude change. Psychological
Review, 1955, fig, 42-55.

 

Patterson, T., and McClure, R. Political advertising:
Voter reaction to televised political commercials.
Citizens' Research Foundation, 245 Nassau Street,
Princeton, N.J., 1973.

 

Patterson, T., and McClure, R. Political campaigns: TV
power is a myth. Psychology Today, 1976, 19, 61.

 

Roberts, D. Nature of communication effects. In W. Schramm
and D. Roberts (eds.), The Process and Effects of Mass
Communication: Urbana University of Illinois Press,
1971.

 

 

Rokeach, M. Beliefs, attitudes, and values. San Francisco:
Jossey-Bass, Inc., 1969.

 

Saltiel, J., and Woelfel, J. Inertia in cognitive pro-
cesses: The role of accumulated information in
attitude change. Human Communication Research,
1975, 1, 333-344.

 

Wiley, D., and Wiley, J. The estimation of measurement
error in panel data. American Sociological Review,
1970, 35, 112-117.

 

 

 

5+1 1

 

     

   

            

HICHIGQN STQTE UN

I ll «lynvllflliyfllfllflf

312931 0