W

  

\

WNWWNW“WWW“WNW

3

-—'l\]_s
Io
_(DJ>-b

 

‘lllllllilllll‘

7N0

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

dissertation entitled

THE MEDIATING ROLE OF VALUES
IN VALUE-EXPRESSIVE ATTITUDES

presented by

Craig Richard Hullett

has been accepted towards fulfillment
of the requirements for a

Ph.D. degree in Communication

,' . l

Mdor prof sssss

Date June 22, 2000

MS U i: an Affirmative Action/Equal Opportunity Institution 0-12771

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

APRLi } 19 320%;

 

 

 

 

 

 

 

 

 

 

 

 

 

moo mm.“

 

THE MEDIATING ROLE OF VALUES IN VALUE-EXPRESSIVE ATTITUDES
By

Craig Richard Hullett

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

DOCTOR OF PHILOSOPHY

Department of Communication

2000

ABSTRACT

THE MEDIATING ROLE OF VALUES IN VALUE-EXPRESSIVE ATTITUDES
By

Craig Richard Hullett

Functional theory predicts that messages designed to change attitudes that are linked to
specific values (value-expressive attitudes) must directly address those values to be
successful. Neofunctional theorists have altered these predictions by arguing that the
desire to conform represents an attitude type (social-adjustive) that opposes the value-
expressive attitude in function, and that either function can be inferred by scores on the
other-directedness dimension of the self-monitoring scale. When integrating these
theoretical approaches, it appears that the particular values of conformity and self-
direction should mediate the relationship between other-directedness and message
perceptions. Overall, the findings of Study 1 indicate that conformity acts as a particular
type of value, and should therefore be considered a potential basis for value-expressive
attitudes. Furthermore, the results are consistent with only the predictions of the original
functional theory. In contrast, the findings from Study 2 indicate that the neofimctional
method of inferring attitude function from measures of other-directedness might have
some value. Other-directedness was related to perceptions of message quality for the two
messages, but the valence of those relationships opposed the directions predicted by
neofunctional theory. Finally, the findings suggest that there may be differences in the

processes caused by the subjects’ involvement with the topic. Subjects may weigh the

value-based arguments against their own values when the topic has high personal

relevance, and rely on more simple heuristics when the topic is of low personal relevance.

ACKNOWLEDGMENTS

I would like to thank my adviser, Dr. Franklin J. Boster. He is a great mentor who
taught me far more than is represented in this document. By example and advice Frank
established in me lofty goals for teaching, research, and service. Most important was his
objective as my major professor. He said he never intended to teach me everything he
knows, but rather enough so that I could learn the rest on my own. Although I’m making
no promise here that I’ll never call on him for advice in the future (and getting rid of me
would be an understandable motivation for that end), I do hope that my future work
shows that he was successful.

I am also indebted to my committee members, Dr. Sandi Smith, Dr. Charles
Atkin, and Dr. John Hunter. First, I owe Chuck and Sandi at least three years of
happiness, because that’s what they provided by bringing me to Michigan State.
Moreover, they provided great advice and challenges when writing this study that
improved its value and expanded my domain of thought. I am grateful to Jack for his
brilliant insights and for strengthening me as a scholar.

I must also thank most of the other members of our faculty for all of their time
and patience with me. Dr. Tim Levine, Dr. Kelly Morrison, Dr. Kim Witte, Dr. William
Donohue, Dr. Jim Dearing, and Dr. Sam Larson all had to endure me for extended
periods of time in research meetings and the classroom. Their efforts opened my mind to
areas of communication I would not have pursued on my own. In the process, they

transferred their excitement about the breadth of the discipline to me.

Also, I thank Dr. Ron Tamborini for three years of excitement. Whether through
guilt, anger, competition, or fun, Ron was always motivated me to remain deeply
involved with the program, the people, and what I was doing. We had a lot of fun, but
that was never at the expense of learning to defend my ideas. Thanks for the fun and
frustration.

To my colleagues I owe even more. Friday football, hanging out at Harper’s, and
poker with the boys (Eric, Tom, Ken, and Matt) were always a blast. Dana and Maria
shared their hearts as well as their time with me when I needed to vent some frustrations,
but were also there to talk business whenever needed. Jenn, and Cynthia could always
sense when it was time to diffuse Diamond Boy’s inner time bomb, and kind enough to
do so. Most important, I gained true friends in my cohort. All of them are irreplaceable,
but Sean really meant the world to me. Thanks for reminding me (almost daily) how fun
life really is (or at least should be).

Finally, and most of all, I must thank Kim. She has given a great deal to this long-
terrn effort of attaining a Ph.D. She was a wonderful critic pointing out flaws or
inconsistencies in my thoughts before they were committed to paper. More important, she
never appeared to lose faith in me during the process. No matter what the circumstances,
she always was a source of strength for me. Most people would not have put up with me
on a daily basis during these last few years (in fact, no one else did), but Kim went
beyond mere tolerance to encourage me throughout the process. Thank you for

everything.

TABLE OF CONTENTS

LIST OF FIGURES ....................................................................... vii
CHAPTER 1

INTRODUCTION ........................................................................ 1
CHAPTER 2

FUNCTIONAL THEORY, SELF

MONITORING, AND ATTITUDE CHANGE ....................................... 3
CHAPTER 3

METHODS ................................................................................ 17
CHAPTER 4

RESULTS ................................................................................. 23
CHAPTER 5

DISCUSSION ............................................................................. 32
APPENDIX A

MESSAGE INDUCTIONS ............................................................. 40
APPENDIX B

STUDY 1 MEASURES .................................................................. 42
APPENDIX C

STUDY 1 CORRELATIONS ........................................................... 44
APPENDIX D

STUDY 2 MEASURES .................................................................. 45
APPENDIX E

STUDY 2 CORRELATIONS ........................................................... 47
LIST OF REFERENCES ................................................................ 48

vi

LIST OF FIGURES
Figure 1 - Predicted Path Model for Conformity Message ............................ 15
Figure 2 — Predicted Path Model for Self-Direction Message ......................... 15
Figure 3 — Obtained Path Model for Conformity Message (Tuition Increase) ...... 25
Figure 4 — Revised Path Model for Conformity Message (Tuition Increase) ........ 25
Figure 5 —- Obtained Path Model for Self—Direction Message (Tuition Increase). . .26

Figure 6 — Revised Path Model for Self-Direction Message (Tuition Increase). . ..26

Figure 7 — Obtained Path Model for Conformity Message (Mentally Ill). . . . . . ..29
Figure 8 — Revised Path Model for Conformity Message (Mentally Ill) ............. 29
Figure 9 - Obtained Path Model for Self-Direction Message (Mentally Ill). . . .30
Figure 10 — Revised Path Model for Self-Direction Message (Mentally Ill). . . ..3O

vii

Chapter 1

INTRODUCTION

The original functional theories of attitudes posited that persuasive messages
would be most effective when directed toward the bases of the attitudes under attack
(Katz, 1960; Smith, Bruner & White, 1956). These original theories listed several
nonexhaustive categories of attitude bases (i.e., “functions”), but provided little direction
for assessing the bases of the attitudes. Perhaps due to this lack of specificity, functional
theory was not pursued seriously for two decades after its introduction.

The recent resurgence in studies of the functional bases of attitudes has been
attributed to advances that allow for the measurement of individuals’ motivational bases
for holding attitudes. Personality variables (Snyder & DeBono, 1985), direct self-reports
(Herek, 1987), and attitude object attributes (Shavitt, 1990) are offered as ways to
determine the functions of attitudes. These alterations to the original theory have led to
the establishment of a “neofunctional” theory (Herek, 1987). Neofunctional theory retains
the basic tenet that persuasive messages must address the function of an attitude to be
successful, but have removed the necessity to assess functional bases. Instead, functions
need only to be inferred from other related measures. One particular approach advocated
by neofunctional theory-—using scores from the self-monitoring scale to infer value-
expressive and social-adjustive attitudes—-has led to data consistent with the main tenet of
neofunctional theory (Snyder & DeBono, 1985). Unfortunately, this approach has failed
to recognize serious flaws in the self-monitoring scale which call into question whether

the construct actually exists (Dillard & Hunter, 1989). These flaws indicate that the self-

monitoring measure is multidimensional, and a careful explication of the theory indicates
that the other-directedness dimension of the scale may be useful for explaining the effects
found in tests of neofunctional theory.

Furthermore, neofunctional theory reconceptualized some of the functions. Most
notably, neofunctional theory posits value-expressive and social-adjustive attitudes as
opposites. Value-expressive attitudes are defined as attitudes that are used to attain some
ideal (i.e., valued) state, whereas social-adjustive attitudes are used to conform with
important reference groups (DeBono, 1987; Lavine & Snyder, 1996). In neofunctional
theory these functions oppose each other, in that an attitude can either be social-adjustive
or value-expressive. Thus, neofunctionalists view these functions as mutually exclusive,
or opposing ends of a single dimension. To value theorists, however, conformity is itself
a value (Schwartz, 1990; Scott, 1965). Thus, neofunctional theory fails to recognize that
social—adj ustive attitudes are a specific category of value-expressive attitudes.
Unfortunately little current research has studied the role of various different values in
mediating the effects of matching messages to inferred attitude flinctions. Because value-
expressive attitudes are generally defined as reflections of the attitude holders’ underlying
values, omitting measures of values represent a serious flaw in current research. The
present studies address the role played by values in mediating the relationship between
self-monitoring (or at least its measure) and attitudes. More specifically, these studies are
designed to test the hypotheses that the values of self-direction and conformity mediate
the relationship between subjects’ levels of other-directedness and perceptions of
message quality when presented with messages designed to change value-expressive and

social-adjustive attitudes.

Chapter 2
FUNCTIONAL THEORY, SELF-MONITORING, AND ATTITUDE CHANGE

The basic arguments of the functional approach are that people hold attitudes for
reasons, and attitudes are not easily changed by arguments that fail to address those
reasons (Katz, 1960). Katz (1960) argued that attitudes serve various needs for people and
that the persuader must address the needs being served to be effective in changing
attitudes. Arguments that do not address the bases of an attitude have little chance to
change the attitude. Moreover, the reasons for the same attitude may vary among people.
Two people might have the same generally positive attitude toward a politician, but one
might primarily hold the attitude because the politician expresses values similar to one’s
own (e.g., equality or independence), whereas the other primarily holds the attitude
because he or she perceives that important others hold a positive attitude about the
politician. When these differences occur, two very different arguments would need to be
constructed to change these attitudes. Hence, without knowing the function an attitude
serves, the persuader cannot effectively design a message to change the attitude.

Initial conceptualizations of the functional bases of attitudes varied only slightly.
At approximately the same time the teams led by Smith (Smith, Bruner & White, 1956)
and Katz (1960; Sarnoff & Katz, 1954) provided similar conceptualizations of the
primary functions attitudes serve. The value-expressive function allows people to feel a
sense of gratification in displaying what they perceive to be important aspects of their
selves. These attitudes are systematically related to the core values people have. Thus, to

change a value-expressive attitude a message must either arouse “some degree of

dissatisfaction with one’s self-concept or its associated values” or “dissatisfaction with
old attitudes as [inconsistent with] one’s values” (Katz, 1960, p. 189).

As Eagly and Chaiken (1993) note, Katz failed to recognize the potential
affiliative benefit of the value-expressive function. Katz did, however, conceptualize this
type of attitude as public (i.e., it is “expressed”). Thus, his definition implicitly includes
the idea that the pleasure derived from holding these attitudes may come from
reinforcement by others holding similar attitudes. Smith et a1. (1956) were more explicit
in their description of this aspect of the value-expressive fiinction within their social-
adjustment function. To Smith and his associates the social-adjustive function is served
when the expression of attitudes serves to promote or preserve relationships with
important others, or break from those relationships no longer considered valuable to the
attitude holder. Although Katz’ definition of value-expressive attitudes can be stretched
to fit both of these functional dimensions, most subsequent researchers have considered
the dimensions separately--where attitudes can be either strongly or weakly based in
values (value-expressive), or have either strong or weak affiliative benefits for the attitude
holder (social-adj ustive). In contrast to value-expressive attitudes one would change a
social-adjustive attitude by convincing the target that holding that attitude will not (or
does not) have affiliative benefits for the attitude holder. As will be discussed later,
however, recent tests of functional theory as it pertains to value-expressive and social-
adjustive attitudes have tended to use messages that address particular values to change or
induce both types of attitudes. Furthermore, one personality variable has been used to
differentiate between holders of each type of attitude. In general, the weaknesses of

current research indicate that only value-expressive attitudes are being studied. Instead of

varying or assessing two types of attitudes, researchers are probably using value-
expressive attitudes that vary in their relationship with two or more underlying values.
Current Research Regarding Value-Expressive and Social-Adjustive Attitudes

Most studies of value-expressive and social-adjustive attitudes are predicated on
Herek’s (1986) neofunctional theory. In Herek’s reconceptualization of the theory the
value-expressive label is “associated with attitudes whose primary motivation results
from some need to affirm one’s sense of self by articulating basic values integral to that
self-concept” (p. 104). Social-adjustive attitudes are “motivated by a need to be accepted
by important others” (Herek, 1986, p. 104). Unlike value—expressive attitudes, social-
adjustive attitudes do not necessarily reflect underlying values, but instead are used to
promote affiliation. That is, social-adjustive attitudes are formed so that the holder can
become a better integrated group member. In general, these conceptualizations are very
close to those offered before. The somewhat subtle departure taken from earlier
conceptualizations is that Herek placed these two dimensions at opposing ends of a single
continuum such that attitudes could be either indicative of the attitude holders’ values
(value-expressive) or reflections of the attitude holders’ perceptions of the attitudes of
others (social-adjustive).

This unidimensional conceptualization led to the use of one personality variable to
indicate both types of attitudes. Herek argued that people with different personality types
would tend to form attitudes for different reasons. Thus, researchers should only need to
identify which personality variables would be associated with particular functions and
then infer attitude function by assessing subjects’ personality types. Herek (1987) and

others (Snyder & DeBono, 1985) argued that self-monitoring differentiates between

people who form attitudes for social-adjustive or value-expressive reasons. In the main,
high self-monitors were expected to hold social-adjustive attitudes because they typically
adjust their behaviors to fit their current social situation (Snyder & DeBono, 1985). In
contrast, low self-monitors would hold value-expressive attitudes because of their
tendency to express their inner thoughts or feelings regardless of social concerns.
Although the basis of a person’s attitude may not be directly accessible for measurement,
the individual difference variable was proposed as an indicator of the bases of subjects’
attitudes.

Placing value-expressive and social-adjustive attitudes at opposite ends of a single
continuum is inconsistent with current research regarding attitude fimctions and common
conceptualizations of values. First, Shavitt’s (1989) research has demonstrated that
specific attitudes can serve multiple functions simultaneously. Some particular attitude
objects (e. g., coffee) may serve only one function (utilitarian), but others (e.g., a fancy
car) may serve multiple functions (utilitarian and social identity) (Shavitt, 1989).

Second, and in conjunction with the first, social-adjustive attitudes may still be
related to the attitude holders’ values and value systems. As Shavitt notes, there is
considerable overlap between Smith et al.’s (1956) conceptualizations of value-
expressive and social-adjustive attitudes. This overlap is a result of the value basis of both
constructs. Rokeach’s (1968) definition of values includes beliefs that a certain “mode of
conduct or end-state of existence is personally and socially preferable” (p. 16). The
social-adjustive fitnction, in its pursuit of establishing and maintaining important
relationships, is consistent with the “socially preferable” end-state in Rokeach’s

conceptualization of values. Similarly, current research indicates that motivations to

conform, have a sense of belonging, and promote social order are key parts of a
“universal value system” (Schwartz, 1992). Thus, the reconceptualization of social-

adj ustive attitudes as lacking a value basis, or as the opposite of a value-expressive
attitude, is implausible. What is more likely is that social-adjustive attitudes are related to
values that motivate affiliative actions. Social-adjustive attitudes would still be value-
expressive, but in a very specific form. In spite of these conceptual shortcomings, most
researchers have still relied on Herek’s method of infen'ing function from subjects’ scores
on the self-monitoring scale.

For the most part the research using the self-monitoring method of inferring
functional type has been consistent with the predictions of neofunctional theory. For
example, using the Self-Monitoring Scale (Snyder, 1974) to divide subjects into High and
Low self-monitor groups, DeBono (1987) predicted that a message advocating a position
based upon the opinions of a majority of the subjects’ peers would match the functional
basis of the subjects who scored high on self monitoring, whereas a different message
advocating the same position because it “reflected an important underlying value” would
match the functional basis of the subjects who scored low in the self-monitoring scale (p.
281). Thus, messages that match the inferred functional basis should be more effective
than messages that fail to match. DeBono found that students’ attitudes toward the
institutionalization of the mentally ill were more affected by a message that matched the
subjects’ level of self-monitoring than by a message that did not match. Similarly, Lavine
and Snyder (1996) used the self-monitoring scale to test the relative strength of matched
versus mismatched messages with the inferred functional bases of their subjects’

attitudes. Their studies of voting attitudes found that the functional match of a positive

message led to increased perceptions of message quality (measured as argument strength)
which, in turn, led to more positive attitudes toward voting. These positive attitudes also
led to increased intentions to vote and, consequently, increased voting behavior. With the
addition of perceptions of message quality mediating attitude change, these two studies
were consistent with the general predictions of neofunctional theory regarding value-
expressive and social-adjustive attitudes.

Bazzini and Shaffer (1995) criticized the methods employed by most of these
studies as assessing attitude shaping rather than attitude change. Bazzini and Shaffer
argued that functional theory is concerned with attitude change and sought to determine
its effectiveness in changing previously established attitudes. They demonstrated that
functionally relevant messages can change important initial attitudes about the criteria
used to select dating partners. They found that counter-attitudinal messages containing
content that matched the attitude function of the subjects more effectively changed
attitudes than messages that were functionally irrelevant. Again, functional relevance of
the messages was assumed by considering value-expressive messages relevant only to
low self-monitors and social-adjustive messages as relevant only to high self-monitors.

Two notable exceptions to the neofunctional approach are provided as
assessments of the structure of value-expressive attitudes. Using Katz’s original theory,
Maio and Olson (1994) derived the hypothesis that strong value-attitude relationships
should only be observed when the attitude serves a value-expressive function. By using a
thought-listing task to differentiate between value-expressive and utilitarian attitudes,
Maio and Olson found that only the group of subjects listing value-expressive attitudes

had a substantial correlation between their values and their attitudes. This pattern also

emerged in a second, experimental study using Schwarz’s (1992) value survey (Maio &
Olson, 1995), lending further credence to the functional perspective regarding attitude
structure.

In sum, the structure of value-expressive attitudes appears consistent with Katz’s
predictions of strong value-attitude relationships. Also, the results of studies assessing the
formation or shaping of value-expressive and social-adjustive attitudes are consistent with
Herek’s predictions regarding high and low self-monitors. Finally, Bazzini and Shaffer’s
study is consistent with the original predictions of Katz, that messages that match the
underlying function of an attitude are more persuasive than messages that fail to match.
What is missing in current research is any attempt to link the research using self-
monitoring and Maio and Olson’s measures of value-attitude relations.

To understand better and map the value-expressive and social-adjustive attitudes,
these two lines of research need to be linked. In general, there is a weak correlation
between self-monitoring and perceptions of message quality when messages match the
function inferred by subjects’ levels of that personality variable. For example, Lavine and
Snyder’s (1996) research found a correlation of 0.35 between self-monitoring and
perceptions of message quality when messages matched. Because message quality was a
mediating variable in the process, they found a weaker correlation between self-
monitoring and attitude. Maio and Olson (1994), however, found similar correlations
between values and attitudes (r = .55 and r = .47 in their two studies regarding attitudes of
nonsmokers toward smoking). Although the sample sizes for all of these studies were
rather small and the correlations were not corrected for measurement error, it is possible

that these actual population correlation between self-monitoring and message quality is

less than 1.0. Thus, there may be another construct mediating that relationship. Given the
focus of functional theory on the importance of matching messages to the values
underlying attitudes, it is consistent with the theory to predict that values serve that
mediating role. More specifically, self-monitoring may affect the types of values people
have. When message recommendations are consistent with these values, perceptions of
message quality will be high and attitudes consistent with message recommendations will
follow. Thus, the observed matching effect between self-monitoring and message quality
would be an indirect effect of messages directly addressing the values associated with
levels of self-monitoring. To predict which values would be related to self-monitoring, it
is important to eXplicate more fully both self-monitoring and that construct’s relationship
with current conceptualizations of values.
Self-Monitoring

In general, self-monitoring was offered as a personality construct representing
people’s abilities to identify and adapt their behaviors to social cues (Snyder, 1974). High
self-monitors are those people who excel at these abilities and are motivated to use these
social skills. Low self-monitors are conceptualized as unconcerned with the perceptions
of others in social situations, and thus unlikely to perceive social cues and adapt their
behaviors. Snyder’s (1974) original measure of self-monitoring was designed to
simultaneously measure five individual differences reflecting these tendencies, and is the
primary measure used in functional theory research for value-expressive and social-
adj ustive attitudes.

Scholars differ regarding which measures of self-monitoring can be used with

confidence (Lennox, 1988). Although Snyder argues that his original Self-Monitoring

10

Scale (Snyder, 1974) and revised version (Snyder & Gangestad, 1986) are valid, more
compelling analyses indicate that neither should be trusted as unidimensional measures of
self-monitoring (Dillard & Hunter, 1989). Dillard and Hunter (1989) found, as did
others, that the Self-Monitoring Scale is actually a multi-dimensional measure, and the
correlations between the factors are zero.

One of the dimensions that continuously emerges in the research is called “Other
Directedness” because the items appear to measure the degree to which subjects adapt
their behaviors and opinions to conform with others. The content of these items generally
indicates a propensity to put forth a front in social situations where the subject is “not
always what I appear to be” or tends to “change [their] opinions in order to please
someone else or win favor” (Dillard & Hunter, 1989, p. 119). These items strongly
parallel Lennox and Wolfe’s (1984) cross-situational variability dimension of their social-
appropriateness scale. The other-directedness dimension strongly reflects the original
conceptualization of the high self-monitor’s ability to adapt behaviors to social demands.
Furthermore, these items are consistent with the neofunctionalists’ views of social-
adjustive attitudes as reflecting the desire to conform.

This dimension of self—monitoring may be related to at least two value
dimensions. First, Schwartz (1992) identified conformity as a general value dimension
that involves the “restraint of actions, inclinations, and impulses likely to upset or harm
others and violate social expectations or norms” (p. 9). This broad value domain is
represented by the specific values of obedience, self-discipline, politeness, and honoring

parents and elders. Because of their desire to restrain themselves from actions that would

11

produce in-group conflict, people who are high in other-directedness should also strongly
value conformity.

Second, Schwartz identified self-direction as a value domain that lies roughly (but
not precisely) opposite to conformity in his circumplex model. This domain is defined as
the desire for “independent thought and action” (Schwartz, 1992, p. 5). These motivations
are represented by the specific values of creativity, freedom, choosing one’s own goals,
curiosity, and independence. People who are high in other-directedness would not hold
these values strongly, because their actions depend so heavily on others in a social
situation. Thus, there should be a negative relationship between other-directedness and
valuing self-direction.

According to functional theory, messages that contain content that is relevant to
these values should be effective in changing subjects’ attitudes when they hold these
values. Extending the model to neofunctional theory, one can infer the types of values
that are likely to be held by subj ects who score high or low on the other-directedness
dimension of the self-monitoring scale. Thus, subjects who score low on other-
directedness should tend to value self-direction, and should therefore be persuaded by
messages that address self-direction values. Subjects who score high on other-
directedness should tend to value conformity, and should therefore be persuaded by
messages that address conformity values.

One of the studies already cited is consistent with these predictions. DeBono’s
(1987) and Bazzini and Shaffer’s (1995) studies could not be used to assess these
relationships because they purposely selected subjects who rated particular values highly,

and used median splits after that selection criterion to determine high and low self-

12

monitors. Thus the relationship between self-monitoring and values cannot be assessed.
Lavine and Snyder (1996), however, do not have this limitation. In Lavine and Snyder’s
(1996) study, the value-expressive message linked message recommendations to the
values of freedom, liberty, and living in a democracy. These values are associated with
both Schwartz’s self-direction value dimension, and its circumplex neighbor,
universalism. Considering the previous discussion of self-direction, one would expect
subjects low in other-directedness to be more persuaded by these arguments than those
scoring high on this dimension. Although Lavine and Snyder’s results are hampered by
their use of the self-monitoring scale in its entirety, their results are consistent with this
prediction. Perceptions of the quality of the value-expressive message were substantially
higher for subjects low in self-monitoring than for those high in self-monitoring.
Similarly, Lavine and Snyder’s social-adjustive message produced effects
consistent with the predictions of values mediating the relationship between self-
monitoring and attitude formation. This message appealed to conformity motives by
telling the subjects that the majority of their peers held attitudes consistent with message
recommendations as well as telling them that the act of voting (their recommended
behavior) would lead to enhanced status and popularity. These latter arguments are more
strongly related to Schwartz’s power dimension, but again this dimension is closely
related to the conformity dimension. Thus, the observed effects should still be consistent
with expectations, but might be weakened by the lack of unidimensionality in the values
addressed. As expected, this social-adjustive message was rated higher in quality by high
self-monitors than by low self-monitors. Thus, both messages can be conceptualized as

appeals to values that are typically associated with subjects’ levels of self-monitoring.

13

More specifically, these results are consistent with what one would expect given subjects’
levels of other-directedness.

Based upon Lavine and Snyder’s (1996) findings, it is also likely that agreement
with message recommendations will lead to increased intentions to enact the behavior
recommended by the message. This prediction is not central to functional theory, but is
consistent with the theory of reasoned action (Fishbein & Ajzen, 1975). A replication of
Lavine and Snyder’s findings in this part of the model coupled with increased
understanding of the initial processes would enhance the value of this adaptation of
functional theory.

The necessary next step is to test these predictions experimentally, rather than
infer them from the results of past studies. These past studies are insufficient to test the
predictions because the messages confounded multiple value dimensions and the
personality measures confounded multiple attributes. Thus, the present investigation will
test two causal models (one for each message). In each figure (see Figures 1 and 2) other-
directedness (OD) is predicted to be positively related to valuing conformity (CONF) and
negatively related to valuing self-direction (SD). The relationships between these values
and perceptions of message quality (MQ) depend on the value-content of the message
induction. Increased perceptions of message quality then leads to increased agreement
with message recommendations (AMR), which subsequently leads to increased intentions

to enact the behavior recommended by the message (BI).

14

Fi ure 1: Predicted Path Model for Conformi Messa e

 

OD

 

 

CNF\MQ
SD/MQ

Figure 2: Predicted Path Model for Self-Direction Message

/\2

OD MQ AMR + BI

 

 

m\
/

Furthermore, the impact of outcome involvement will be assessed. Involvement is

/\g

a construct that is important in message processing. Two popular theories of attitude
change-the elaboration likelihood model (Petty & Cacioppo, 1986) and the heuristic-
systematic model (Chaiken, l987)--posit that subjects are motivated to do more message-
relevant thinking about issues that are perceived as personally important or relevant.
Lavine and Snyder (1996) argue that function-matching effects are only likely when
subjects engage in this message-relevant thinking, because subjects must process the
evidence provided in the messages to be influenced by the match of the evidence to the
function. Thus, the use of tOpics with which the subjects are involved is important for
matching effect to occur. Furthermore, a meta-analysis found that using topics where the
outcome was relevant to subjects tended to enhance the effects of messages containing
strong arguments (Johnson & Eagly, 1989). Argument strength is already being measured

in the present investigation (message quality). Because of its importance in determining

15

message processing, the effects of argument strength, and the potential for variance in

involvement for various topics, outcome involvement must also be assessed.

l6

Chapter 3
METHOD

Overview

Two studies were conducted to test the models. The studies were identical in most
procedures and methods, differing primarily in attitude domain and measures of the
perceived content of the message inductions. The first study focused on tuition increases
and the second study focused on institutionalization of the mentally ill as their respective
attitude domains. In each study, participants’ scores on the other-directedness dimension
of the self-monitoring scale, responses to Schwartz’ values inventory, and initial attitudes
were measured. One week later, the subjects were assigned randomly to conditions in
which they received messages arguing in favor of increasing tuition (or institutionalizing
the mentally ill, depending on the study) for either conformity or self-direction reasons.
The messages indicated that the subjects would have the opportunity to vote on the issues
addressed in the messages in the near future. After reading the assigned message, subjects
then answered items about the content of the message, message quality, attitude toward
the topic, involvement with the topic, and intention to vote in favor of the position
advocated by the message.
Subjects

A total of 287 subjects participated voluntarily in the two studies. These subjects
were recruited from undergraduate classes at Michigan State University, and their
responses will remain confidential. In Study 1 (Tuition) 83 of the 144 (58%) subjects
were female, the average age was 20.43 years (SD = 2.04), 80% were Caucasian, and

15% were African-American. In Study 2 (Mentally Ill) 96 of the 143 (67%) subjects were

17

female, the average age of the subjects was 20.31 years (SD = 1.71), 80% were
Caucasian, and 12% were African-American.
D_es_ign

Both studies were designed as pretest-posttest comparison groups with the first
one-half of the subjects receiving a conformity message, and the other one-half of the
subjects receiving a self-direction message.
Procedure

Initially, the subjects’ other-directedness, values, and attitudes toward the topic
were measured using self-report. Subjects were informed that they were participating in a
study in which they would be discussing some important state and campus issues, and
that they must return the next week to engage in the discussions. One week after the
initial surveys were completed, subjects were presented with the persuasive messages--
either a conformity or self—direction message advocating either increasing tuition (Study
1) or institutionalizing the mentally ill (Study 2). Subjects were informed that they
needed to read the information carefully and answer some questions about the
information before the discussions began. The questions in the survey included items
designed to measure the effectiveness of the inductions, the subjects’ perceptions of
message quality, involvement with the attitude topic, attitudes toward the topic, and
intentions to vote in favor of the position advocated by the message. After these measures
were completed, subjects were debriefed and released from the study.
Messages

Two messages (one conformity message and one self-direction message) were

designed for each of the two studies (see Appendix A). Content from the Schwartz values

18

inventory was used in the messages to ensure a match between message content and the
values comprising the dimensions of interest to this study. Thus, the conformity message
associated the values of politeness, self-restraint, and getting along with others with the
positions advocated within the messages. F urthennore, the conformity messages
contained information stating that the majority of the public and the subjects’ peers
supported the positions advocated by the messages. Likewise, the self-direction message
associated the values of self-sufficiency, self-reliance, ability to choose one’s own goals,
and independence with the message recommendations.

These messages were pretested for content using four items to assess the subjects’
perceptions of conformity and self-direction content (see Instrumentation). The pretest
used 31 subjects who each responded to one message regarding tuition increases and one
message regarding institutionalization of the mentally ill. Subjects received both
functional arguments (one for each attitude domain) to ensure that their responses were
not affected by multiple exposures to the same fundamental argument basis. Furthermore,
the order of the attitude domain presented was randomized to eliminate ordering effects.
Nineteen (61%) of the subjects were female. Twenty-six were Caucasian, two were
African-American, one was Hispanic, and two failed to indicate their race. Subjects’
responses to the four items representing each construct were indicated on seven-point
Likert Scales and summed to form indices. Pretests indicated that the messages were
effective. For the study of attitudes toward increasing tuition, subjects perceived the
conformity message to have higher conformity content (M = 19.75, SD = 6.78) than the
self-direction message (_M_ = 10.67, S_12 = 6.32) [1129) = 3.85, p < .05, r = .58). Subjects

also perceived the self-direction message to have higher self—direction content (M =

19

20.21, SD = 6.41) than the conformity message (M = 7.94, SM: 4.25) [t (28) = 6.26, p_<
.05, r = .76]. The messages did not differ in perceptions of argument strength for the
conformity message (M = 24.00, _S_D_ = 6.81) and self-direction message (M = 23.94, SD
= 9.23) [t (29) = .02, p > .05, r = .00]. In the study of attitudes toward institutionalizing
the mentally ill, subjects perceived the conformity message to have higher conformity
content (M = 17.50, S1; = 4.24) than the self-direction message (M = 9.69, _S_I_) = 4.90) [t
(28) = 4.69, p < .05, r = .66). In the same study, subjects perceived the self-direction
message to have higher self-direction content (M = 23.19, SD = 5.38) than the conformity
message (M = 12.40, SD = 7.24) [t (29) = 4.73, p < .05, r = .66]. Although perceived
argument strength for the conformity message (M = 15.94, SQ = 7.74) and the self-
direction message (M = 18.07, E = 6.25) were somewhat lower than in the other study,
they did not differ within the study [t (29) = .84, p > .05, r = .15]. Thus, all four messages
were used for the present studies because they varied successfully the intended content
without varying message strength.
Instrumentation

Other-Directedness. The five items identified by Dillard and Hunter as comprising
the Other Directedness measure were used to measure this variable. Subjects responded to
the items using a seven-point Likert-type Scale. Higher scores indicate higher levels of
other-directedness.

Value Dimensions. Schwartz’s (1992) universal values scale was used to assess
subjects’ conformity and self-direction values. The 56-item scale was used in its entirety,
representing each of the 11 values identified by Schwartz as “universal.” The inclusion of

these other values allowed for the assessment of any confounding effects that could have

20

arisen if the messages accidentally included references to values other than conformity
and self-direction. Four items measured conformity and six items measured self-direction.
Each item was presented as a value, and the subjects rated the importance of each value
“As a guiding principle in my life” on an ll-point scale ranging from 0 (not at all
important) to 10 (of supreme importance) (Schwartz, 1992, p. 17).

Self-Direction Content of Message. Four items were written to assess the
subjects’ perceptions of the self-direction content of the messages. Subjects indicated
their agreement with statements about the content of the messages on seven-point Likert
Scales. Scores were summed so that higher values indicate greater perceptions of self-
direction content.

Conformity Content of Message. Four items were written to assess the subjects’
perceptions of the conformity content of the messages. Subjects indicated their agreement
with statements about the extent to which the messages contained conformity content on
seven-point Likert Scales. Higher scores indicate greater perceptions of conformity
content.

Message Quality. Perceived message quality was measured by Lavine and Snyder
(1996) as the strength of the arguments presented by the messages. They used items such
as “I found the material to be convincing” to form the measure (Lavine & Snyder, 1996,
p. 588). Although their entire measure was not available, another reliable measure of
argument strength was presented by LaFrance (1998). In her study, argument strength
was measured using seven semantic-differential items (believable/unbelievable,
compelling/not compelling, convincing/unconvincing, logical/illogical,

plausible/implausible, reasonable/unreasonable, and sound/not sound). Six of these terms

21

were used for the present study, but split into separate statements to which the subjects
indicated their level of agreement or disagreement ~with seven-point Likert Scales. Higher
scores indicate greater perceptions of message quality.

Agreement with Message Recommendation and Attitude Change. Attitudes were
measured at the pretest and posttest using the same items. The items measured the
subjects’ attitudes toward the two topics covered by the two studies. In Study 1 six items
were used in the pretest and posttest to measure the subjects’ attitudes toward increasing
tuition, and in Study 2 six similar items were used in the pretest and posttest to measure
the subjects’ attitudes toward institutionalizing the mentally ill. The posttest measures are
used as indices of agreement with message recommendations. Subtracting the pretest
measures from the posttest measures allows for assessment of whether the messages were
successful in changing attitudes. For the posttest measure, items were summed such that
higher scores (on seven-point Likert Scales) indicate greater agreement with message
recommendations. For attitude change, items that were retained from the pretest were
subtracted from the posttest measures, and these differences were summed so that higher
scores indicate greater attitude change in the direction advocated by the message.

Involvement. Involvement was measured with four items designed to assess the
extent to which subjects felt the outcome of the vote would affect them. Subjects
responded to four items using seven—point Likert Scales. Scores were summed so that

higher scores indicate greater involvement.

22

Chapter 4
RESULTS
Study 1 (Increasing Tuition)
Test of the Measurement Model
Confirmatory factor analyses (Hunter & Gerbing, 1982) were used to test the
validity of the hypothesized factor structure of the items, and assess the reliability of the
measures. All of the items retained for this study and their factor loadings can be found in
Appendix B. The following results indicate the descriptive statistics of the measures, and
reliabilities indicated by standardized item alpha. All scale means are on a scale from one
to seven, except for the value dimensions which range from zero to ten.
Other-Directedness. Four items were retained as indicators of other-directedness
(M = 3.53, SD = 1.17, or = .67).
Value Dimensions. Three items were retained measuring conformity (M = 8.18,

__I_)_ = 1.48, or = .71), and all six items were retained as indicators of self-direction (M =
8.67, _S_D = .99, or = .79).

Self-Direction Content of Message. Four items were retained as indicators of self-
direction content (M = 4.30, SD = 1.90, or = .94).

Conformity Content of Message. Four items measuring conformity content were
retained (M = 4.11, SD = 1.53, a = .87).

Message Quality. Five items were retained as measures of message quality (M =
4.41, _S_D=1.17, or = .87).

Agreement with Message Recommendations. Six items measuring the posttest

attitudes of the subjects were retained as valid indicators (M = 4.15, S_D = 1.52, or = .96).

23

Attitude Change. Because only five items from the pretest measure of attitudes
were retained, only five could be used for the measure of attitude change (M = .31, S_D =
1.35, g = .93).

Involvement. Four items measuring involvement were retained (M = 4.40, §I_) =
1.60, or = .80).

Behavioral Intentions. Six items were retained as indicators of behavioral
intentions (M = 3.89, S1; = 1.70, or = . 96).

Induction Check

The results indicate that the messages were successful in conveying the type of
information intended. The conformity content was stronger in the conformity message (M
= 4.68, S_D = 1.09) than in the self-direction message (M = 3.54, S_D = 1.69) [t (142) =
4.84, p < .01, r = .37]. Also, the self-direction content was stronger in the self-direction
message (M = 5.62, S_D = 1.11) than in the conformity message (M = 2.98, SD = 1.59) [t
(142) = 11.52, p < .01, r = .69]. The differences were not as strong as observed in the
pretest, but were still substantial. Thus, the messages were retained because they were
successful in varying functional content.

Test of the Model for the Conformity Message

In the conformity message condition, the data were not consistent with the
hypothesized model (See Figure 3, and see Appendix C for correlations among measures
used in the path analyses). It was hypothesized that other-directedness would be related
positively to conformity and negatively related to self-direction, but these path

coefficients were not substantial. Corrected correlations indicate that other-directedness

24

Figure 3: Obtained Path Model for Conformity Message (Tuition Increase)

 

 

CONF
.13 y .34(.17)
OD MQ .26(.12) AMR .95 (.02) BI
05% A7)

SD

 

x: (9) = 8.47, p = .49, n = 72; Standard errors in parentheses
was unrelated to valuing conformity [r’ (70) = .13, p > .05] and valuing self-direction [r’
(70) = .05, p_ > .05] in the conformity condition. Thus, the model was rejected.

A revised model was tested that was consistent with the data. As Figure 4
illustrates, the data were consistent with a revised model positing that valuing conformity
was related positively with perceptions of message quality [p (70) = .34], and valuing
self-direction was related negatively to perceptions of message quality [p (70) = -.23].
Perceptions of message quality were related substantially to agreement with message
recommendations [p (70) = .26], which were related strongly to behavioral intentions [p
(70) = .95]. Residuals were all explainable by sampling error, and the overall test of the
model’s residuals was insignificant. Thus, the revised model was not rejected for the
conformity message.

Figure 4: Revised Path Model for Conformig Message (Tuition Increase)
CONF

 

.34 (.12)

MQ .26(.12) AMR .95 (.02) BI

 

 

-.23 (.12)
SD

 

x2 (5) = 3.07, p = .69, n = 72; Standard errors in parentheses

25

Figure 5: Obtaifined Path Model for Self-Direction Messa e Tuition Increase
CONF

-.26()2/ -.O6 (.12)
OD MQ
-16(\1) M1)
SD

 

.66(.06) AMR .94(.02) BI

 

 

 

x2 (9) = 11.91, p = .22, n = 72; Standard errors in parentheses

Test of the Model for the Self-Direction Messa e

 

Similar predictions were tested in the self-direction message. As in the previous
condition, the model was not consistent with the data. As is evident in Figure 5, other-
directedness was unrelated to valuing conformity [r’ (70) = -.26, p > .05] and self-
direction [1’ (70) = -.l6, p_ > .05] (see Appendix C for remaining correlations). Thus, the
data were not consistent with proposed model, and the model was rejected.

A revised model was obtained that was consistent with the data. Figure 6
illustrates a model in which valuing self-direction leads to increased perceptions of
message quality [p (70) = .28] when receiving a self-direction message. These
perceptions were related strongly to agreement with the message recommendations [p
(70) = .66], and agreement with message recommendations was related subsequently to
behavioral intentions [p (70) = .94]. Again, residuals were within the range one would

expect due to sampling error, and the overall test of the model’s residuals was

insignificant. Thus, the revised model was not rejected.

Figure 6: Revised Path Model for Self-Direction Message (Tuition Increase)
SD .28 (.10) MQ .66 (.06) AMR .94 (.02) BI

 

 

 

 

x3 (3) = 1.73, p = .63, n = 72; Standard errors in parentheses

26

Study 2 (Institutionalizing the Mentally 111)

Test of the Measurement Model

Confirmatory factor analyses (Hunter & Gerbing, 1982) were used to test the
validity of the hypothesized factor structure of the items, and assess the reliability of the
measures. All of the items retained for this study and their factor loadings can be found in
Appendix C. The following results indicate the descriptive statistics of the measures, and
reliabilities indicated by standardized item alpha. All scale means are on a scale from one
to seven, except for the value dimensions which range from zero to ten.

Other-Directedness. Three items were retained as indicators of other-directedness
(M = 3.56, _S_D = 1.38, or = .68).

Value Dimensions. Four items were retained measuring conformity (M. = 8.07,
SD = 1.23, or = .69), and all six items were retained as indicators of self-direction (M =
8.68, SD = .88, or = .71).

Self-Direction Content of Message. Four items were retained as indicators of self-
direction content (M = 4.14, SD = 1.86, or = .94). ’

Conformity Content of Message. Four items measuring conformity content were
retained (M = 3.96, S_D = 1.66, or = .90).

Message Qualig. Four items were retained as measures of message quality (M =
4.66, S_D = 1.13, or = .76).

Agreement with Message Recommendations. Six items measuring the posttest

attitudes of the subjects were retained as valid indicators (M = 5.07, _S_D = 1.42, or = .94).

27

Attitude Change. Because only four items were retained fi'om the pretest measure
of attitudes, only those four were subtracted from similar items in the posttest to obtain
the measure of attitude change (M = .047, S_D = 1.27, g = .92).

Involvement. Four items measuring involvement were retained (M = 3.52, SD =
1.16). These items had low reliability (or = . 73).

Behavioral Intentions. Five items were retained as indicators of behavioral
intentions (M = 4.65, SD = 1.34). These items comprised a scale with good reliability (or
= . 89).

Induction Check

The results indicate that the messages were successful in conveying the type of
information intended. The conformity content was stronger in the conformity message (M
= 4.70, SD = 1.22) than in the self-direction message (M = 3.22, SD = 1.71) [t (142) =
6.01 , p < .01 , r = .45]. Also, the self-direction content was stronger in the self-direction
message (M = 5.26, _S_D = 1.26) than in the conformity message (M = 3.01, SD = 1.68) [t
( 142) = 9.08, p < .01, r = .60]. The strength of these effects was similar to the effects
observed in the pretest. Thus, the messages were considered successfirl in varying
functional content.

Test of the Model for the Conformity Message

Path analyses led to rejection of the hypothesized model (see Figure 7, and see
Appendix E for correlations among variables used in path analyses). It was hypothesized
that other-directedness would be positively related to conformity and negatively related to

self-direction, but these path coefficients were not substantial.

28

 

Fi ure 7: btained Path Model for onformi Messa e Mental] Ill

 

 

 

CONF
.01 y -.19 (.24)
OD MO .62 (.10) AMR .86 (.02) BI
-09(.\7) M3)

SD

 

x2 (9) = 13.68, p = .13, n = 72; Standard errors in parentheses

Clearly, other-directedness was not a substantial cause of valuing conformity [p (70) =
.01] or self-direction [p (70) = -.09]. Furthermore, these values did not appear to be
substantially related to perceptions of message quality. Thus, the model was rejected.

A revised model was obtained that was consistent with the data. Figure 8
illustrates that when receiving a conformity message, other-directedness was negatively
related to perceptions of message quality [p (70) = -.38]. Perceptions of message quality
were strongly related to agreement with message recommendations [p (70) = .62] which,
in turn, was strongly related to behavioral intentions [p (70) = .86]. The exogenous path
in this model is drastically different from what would be obtained from the original
hypotheses. It was hypothesized that other-directedness would have an indirectly positive
relationship with perceptions of message quality in this condition. The converse,
however, was found. Because the revised model had predicted correlations with sampling
error of those obtained and the overall test of the models’ residuals was insignificant, the
model was not rejected.

Figure 8: Revised Path Model for Conformity Message (Mentally Ill)

OD-.38(.15) MQ .62(.10) AMR .86(.05) BI

 

 

 

 

x2 (3) = 2.23, p = .53, n = 72; Standard errors in parentheses

29

Figure 9: Obtained PaLModel for Self-Direction Message (Mentally Ill)

CONF
49/ \36 (.23)
OD MQ
~0me M3)

SD

.53 (.13) AMR .83 (.06) BI

 

 

 

x2 (9) = 13.13, p = .16, n = 71; Standard errors in parentheses

Test of the Model for the Self-Direction Message

This model was rejected for the same reasons found in the previous model. Figure
9 illustrates the lack of any substantial relationship between other-directedness and
valuing either conformity [p (69) = -.26] or self-direction [p (69) = -.06]. Also, these
values did not predict perceptions of message quality. Thus, the model was rejected.

A revised model was found that removed the values as mediating variables in the
process (see Figure 10). Other-directedness was positively related to perceptions of
message quality in this condition [p (69) = .35]. The remaining paths from message
quality to agreement with message recommendations [p (69) = .53], and from agreement
with message recommendations to behavioral intentions [p (69) = .83] were consistent
with previous models. Like the conformity message, the relationship between other-
directedness and perceptions of the quality of the self-direction message was the inverse
of the direction expected. The residuals for this model were again small, and the overall
test of the model’s residuals was insignificant. Thus, the model was not rejected.

Figure 10: Revised Path Model for Self-Direction Message (Mentally Ill)

OD .35(.16) MQ .53(.11) AMR .83(.06) BI

 

 

 

 

x2 (3) = 1.36, p = .72, n = 71; Standard errors in parentheses

30

Post hoc Analyses

Attitude Change

As is indicated by mean attitude change in each study, the messages did not
appear to be effective overall in changing attitudes. In Study 1, attitude change (M = .51,
_S_D = 1.46) was substantially greater than zero only for the self-direction message [I (71)
= 2.99, p_ < .05]. For the conformity message attitude change (M = .10, SD = 1.21) did not
differ from zero [I (71) = .72, p > .05]. In Study 2, attitude change failed to differ from
zero (M = -.01, SD = 1.34) for the conformity message [t (71) = .09, p_ > .05], and also
failed to differ from zero (M = .11, _SD = 1.21) for the self-direction message [I (70) =
.76, p_ > .05]. Based on the general failure of the messages to actually change subjects’
attitudes, the previous results are limited to attitude shaping rather than attitude change.
Involvement

The differences in obtained models between the two studies necessitated some
investigation of the differences in outcome involvement for the two attitude domains. As
is indicated in the summaries of the measures for each study, subjects’ outcome
involvement was substantially higher for the increased tuition topic (M = 4.40, SD =
1.60) than for the mentally ill topic (M = 3.52, S_D = 1.16) [t (285) = 5.37, p < .01, r =
.30]. These differences may have led to some important differences in message
processing, and thus the effectiveness of the message content in addressing the fimctional

bases of subjects’ attitudes.

31

Chapter 5
DISCUSSION

Overall, the tests of the proposed models were inconsistent with the predictions of
the investigation. Other-directedness was consistently independent of valuing conformity
and self-direction. Thus, these values could not serve a mediating role in the proposed
process of attitude change. There were some interesting findings within the studies,
however, that merit further discussion. This final section discusses the findings within
each study, potential reasons for the differences between the studies, implications for the
study of functional theory, limitations of the investigation, and directions for future
research.
Sam

The results of Study 1 indicate that persuasive messages are effective to the extent
that their value content matches the values of the message recipients. The perceived
quality of the conformity message was related positively to the subjects’ valuing of
conformity. Likewise, the perceived quality of the self-direction message was positively
related to the subjects’ valuing of self-direction. These perceptions of message quality
were related subsequently to greater agreement with message recommendations and
increased intentions to enact the behaviors advocated by the messages. These results were
consistent with the fundamental claim of functional theory regarding value-expressive
attitudes.

In contrast, the claims of neofunctional theory were not consistent with the data.
Specifically, other-directedness did not belong in the process model because it failed to

predict the extent to which subjects valued conformity and self-direction. Thus, the use of

32

 

the self-monitoring scale (or at least one primary dimension) was not effective in
predicting the effectiveness of the messages.

Finally, the data indicate that treating conformity and self-direction as two distinct
values is warranted. Although these two values appear to lie toward opposite poles in
Schwartz’ circumplex, they did not prove to be second-order unidimensional in this
investigation. For both messages the two dimensions were statistically independent of
each other. Thus, at least when following Schwartz’ directions for measuring these
dimensions, the two dimensions should continue to be treated as separate values.
Moreover, because of their usefulness in predicting the effects of self-direction and
conformity messages, both should be considered potential bases for value-expressive
attitudes.

Mil

The subjects did not consider their own values when considering the strength of
the messages advocating institutionalizing the mentally ill. Instead, higher scores on the
other-directedness scale were directly associated with considering the conformity
message to be of lower quality. Also, higher scores on other-directedness were directly
associated with higher ratings of the self-direction message. These findings not only fail
to include values as mediating variables, but also oppose the relationships initially
predicted.

One potential explanation is that the subjects may have been more focused upon
how their votes would impact the mentally ill rather than themselves. The self-direction
message could have been seen as consistent with the high other-directedness subjects’

opinion of fitting in with others. That message--as opposed to the conformity message--

33

explicitly states that the ultimate goal of institutionalization of the mentally ill is to
reintegrate the patients with society. The subjects may have perceived the implicit
argument that release into society presumes that the patients are able to conform with the
rules of society. Gains in this ability might be considered important to people who follow
the same rules in their own interactions. Therefore, people who tend to be other-directed
might see this type of argument as reasonable. Likewise, the subjects may have noticed
that such intentions and arguments were lacking in the conformity message. The
conformity message would then be considered of lower quality to the extent that it failed
to align with the subjects’ standards of others’ social behaviors. Thus, if the subjects were
focused on the future interactions of the mentally ill rather than conforming themselves
with others’ attitudes, the relationships observed between the perceptions of message
quality and other-directedness would be expected.

Another interpretation is plausible. Because the subjects’ values were not related
to their perceptions of message quality, it is possible that attitudes toward the mentally ill
are not value-expressive attitudes. Attitudes can have multiple functions (Shavitt, 1989),
and are not necessarily based in a value-expressive function. If these attitudes are not
value-expressive, they should not be expected to act in the same way as value-expressive
attitudes. The present investigation failed to determine whether these attitudes are value-
expressive, so this alternative explanation cannot be evaluated. Future investigations
should attempt to make this determination beforehand. That way researchers can avoid
the problem of nonfalsifiability in claiming that attitudes that fail to follow the theoretical

process are not the types of attitudes specified by the theory.

34

Differences Between the Two Studies

The primary differences between the two studies lie in the exogenous paths. Not
only did the variables determining message quality differ, but the valence of the paths
also reversed between the two studies. The subjects’ levels of outcome involvement
provided the only measured differences between the studies. The potential influence of
this variable is thus explored as an explanation for the observed differences.

As was found in the post hoc analyses, subjects’ outcome involvement was
substantially higher for the increased tuition topic (Study 1) than for the mentally ill topic
(Study 2). Finding the topic personally relevant in Study 1 may have motivated the
subjects to consider more fully the evidence in the message (Chaiken, 1987). With this
increased attention to the message, the subjects would have been more likely to process
systematically the messages and notice the links between the values espoused in the
messages and their own values. Involvement may be required for subjects to attend to the
similarity or discrepancy between the values cited in a persuasive message and their own
values. Consequently, high outcome involvement would be necessary for value-
expressive messages to be effective.

In Study 2, these effects were unlikely because subjects would not have been as
motivated to process the message against their own values. The previous discussion of
Study 2 indicates a different concern regarding the outcome of the votes. In Study 1, the
outcome was likely to be directly relevant to the subjects, whereas the outcomes would
only be indirectly relevant to the subjects in Study 2. Thus, the subjects may not have felt
the need to compare the arguments of the message to their own values, but only to their

standards for social behavior. If this difference in outcome involvement was operating,

35

the links between the values cited in the messages and the values held by the subjects
would be missed.

The above explanation implies that if subjects in Study 2 were not processing
systematically, then they might have been processing heuristically. Perhaps conformity
with social norms operated as a heuristic for subjects high in other-directedness in Study
2. While the data are consistent with the argument that subjects were processing messages
systematically in Study 1, the data do not necessarily indicate heuristic processing in
Study 2. It seems that the perception of an implied argument would require some
cognitive effort in scrutinizing a message that is at least as substantial as the effort
required to elaborate the explicit claims of a message. The low level of involvement
observed in Study 2, however, suggests that subjects would not be motivated to put forth
such effort. It is possible, however, that this heuristic is highly accessible (especially for
people who are highly other-directed) and, therefore, able to exert some persuasive force
without much effort (Chaiken & Eagly, 1983). Unfortunately, no measures of the
accessibility of this heuristic were used for this study, so this post hoc hypothesis cannot
be evaluated.

Implications for Functional Theory

A primary goal of the present investigation was to integrate the original functional
theory with neofunctionalist perspectives. The results do not support that integration.
Other-directedness was generally independent of the subjects’ values across all four
conditions, indicating that the values do not mediate the path between other-directedness
and message perceptions. Consequently, at least according to the rationale of this

investigation, the two perspectives cannot be integrated.

36

Moreover, the results of these two studies make it difficult to choose one
functional theory over the other. In Study 1, the data were consistent with the original
functional theory, because perceptions of message quality were directly linked to the
subjects’ underlying values. The results of Study 2, however, appear more consistent with
neofunctional theory in that the message perceptions were directly linked to other-
directedness. It is important to note that the personality variable was only proposed as an
indicator of the types of attitudes a person would be likely to hold. If not related to
specific values, the personality variable was supposed to at least predict that the quality of
value-based message would be negatively associated with other-directedness. Instead, the
perceived quality of the conformity message (which is more representative of a
neofunctionalists’ social-adjustive message) was the negatively related to other-
directedness, whereas the opposite results were found for the self-direction message
(which is more like the neofunctionalists’ value-expressive message). Thus, the results
were just the opposite of what neofunctional theory would predict. Consequently, the
more traditional approach seems more effective in predicting subject responses.

Finally, the differences between the studies underscore the importance of
establishing the functional bases of attitudes before testing the underlying processes. The
data from both studies might indicate that messages addressing the reasons underlying
subjects’ attitudes will be more effective than messages that fail to address these bases.
The matching effect predicted by both functional approaches cannot be inferred, however,
unless it is assumed that the pattern of data in Study 2 is explained by implicit arguments

in the messages. The need to make such an assumption could be avoided by first

37

determining the typical function being served within a particular attitude domain (Shavitt,
1990).
Limitations and Directions for Future Research

In addition to the limitations and future directions recommended above, it seems
important to gain some sense of how and when value-expressive attitudes are developed.
This investigation did not begin with an assessment of the functions served by the
attitudes studied, and therefore cannot conclude that value-expressive attitudes were
changed. Instead, it may be that attitudes in Study I became value-expressive because of
the arguments presented. That is, the subjects general attitude toward increasing tuition
may have remained constant, but the reasons for the attitude may have shifted to agree
with the reasons presented in the messages. Future investigations should not only assess
the functional bases of the attitudes to better determine underlying processes, but should
also attend to the possibility that persuasive arguments change the fimction without
changing the basic valence of the attitude. If people can maintain attitudes but shift
functions, it would be important to understand whether the process fimctions as a form of
resistance. People might be able to protect attitudes and behaviors for consistency
purposes by simply changing the underlying reasons for them.

Finally, future research should attempt to determine how (or if) these processes
are addressed in interactions. This experimental design lends itself more to mass media
campaigns than interpersonal applications. Although the most practical application of
fimctional research may be in political and advertising venues, it would still be useful to
know about these processes in relationships. To date, no published research has

determined whether people attempt identify the functions being served by a particular

38

 

attitude, and then craft messages that address the functions. Some exploratory
investigation of everyday interactions may add more insight to these processes.
Conclusion

The abandonment of functional theory in favor of the short-cuts provided by
neofunctional theory is not warranted. The kindest interpretation of these findings
indicate that one dimension of the self-monitoring scale can be useful in predicting
responses to value-expressive messages. Less kind (and probably more accurate)
interpretations indicate that the neofunctional theory provides no basis for explaining or
measuring the processes underlying value-expressive attitudes. Functional theory,
however, also may have failed to explain fully the results. It may be that the attitude
domain in Study 2 was not typically value-expressive, but it is at least equally plausible
that involvement played a role in the differences between the two studies. Thus, while
there does not seem to be any evidence supporting the integration of functional theory
with neofunctional theory, perhaps an integration of the original theory with the heuristic-

systematic model is warranted.

39 I

APPENDIX A

MESSAGE INDUCTIONS

Study 1: Conformity Message

Currently, the Michigan State University Board of Trustees is considering a
proposal to increase tuition at Michigan State University. The tuition increase would be
dedicated to hiring more faculty members for each department at the university. The
increase would amount to approximately $200 per semester for each student enrolled at
Michigan State University. The Board’s decision about the proposal will depend on the
percentage of students who vote in favor or against the tuition increase.

A recent survey of students at Michigan State found that the majority (72%) are in
favor of increasing tuition to get more faculty. It also found that the Michigan State
students in favor of increasing tuition value “politeness” and “self-restraint” highly.
These students also tend to fulfill their obligations, and “get along” well with their peers.
Thus, people who vote “yes” for the proposal should be the types of people who get along
well with others. So, most students should support the tuition increase.

Study 1: Self-Direction Message

Currently, the Michigan State University Board of Trustees is considering a
proposal to increase tuition at Michigan State University. The tuition increase would be
dedicated to hiring more faculty members for each department at the university. The
increase would amount to approximately $200 per semester for each student enrolled at
Michigan State University. The Board’s decision about the proposal will depend on the
percentage of students who vote in favor or against the tuition increase.

If tuition is increased, students at Michigan State will have more classes to choose
from, which increases their ability to choose their own educational goals. Students will
have more freedom in selecting the courses they take to fulfill requirements for their
major, and be more self—reliant in designing their own curriculum according to their own
desires and interests. Thus, the tuition increase will help students become more
independent. So, most students should support the tuition increase.

Study 2: Conformity Message

Currently, there is a petition to put an issue about institutionalizing the mentally
ill on the Michigan state ballot. This initiative would allow psychiatrists to place
Michigan residents who are diagnosed with severe mental illnesses into temporary
housing facilities where they receive 24-hour attention. Patients with mild forms of
depression, obsessive-compulsive disorder, and attention deficit disorder would not be
subject to institutionalization. A second opinion by another psychiatrist would be
required within 24-hours to determine whether the patient needs to remain in the facility.
If the bill receives a majority of “yes” votes in November, it will allow psychiatrists to

40

immediately place people with these illnesses into institutions that can house and care for
the patients.

A survey of registered voters found that almost three-fourths (74%) of the people
support this initiative. Similarly, last week’s survey of students at Michigan State found
that the majority of young-adults of voting age (72%) are in favor of institutionalizing
peOple with severe mental illness. Students in favor of this proposal value “politeness”
and “self-restraint” highly. These students also tend to fulfill their obligations, and “get
along” well with their peers. Thus, people who vote “yes” for the initiative should be the
types of people who get along well with others.

Study 2: Self-Direction Message

Currently, there is a petition to put an issue about institutionalizing the mentally
ill on the Michigan state ballot. This initiative would allow psychiatrists to place
Michigan residents who are diagnosed with severe mental illnesses into temporary
housing facilities where they receive 24-hour attention. Patients with mild forms of
depression, obsessive-compulsive disorder, and attention deficit disorder would not be
subject to institutionalization. A second opinion by another psychiatrist would be
required within 24-hours to determine whether the patient needs to remain in the facility.
If the bill receives a majority of “yes” votes in November, it will allow psychiatrists to
immediately place people with these illnesses into institutions that can house and care for
the patients.

While institutionalized, these patients could learn to be more self-sufficient.
Patients would also gain the skills to become self-reliant. While under their doctor’s care,
these patients could gain the ability to choose their own goals in life. Then, the patients
would be given the freedom to pursue their goals when they are released fiom the facility.
Thus, institutionalization of people with severe mental illnesses would ultimately help to
increase the independence of these people.

41

APPENDIX B

STUDY 1 MEASURES

Other-Directedness

In different situations and with different people, I often act like very different
persons.

In order to get along and be liked, I tend to be what people expect me to be
rather than anything else

I’m not always the person I appear to be

I may deceive people by being fiiendly when I really dislike them

I would not change my opinions (or the way I do things) in order to please
someone or win favor

Conformity Value

POLITENESS (courtesy, good manners)

SELF-DISCIPLINE (self-restraint, resistance to temptation)
HONORING OF PARENTS AND ELDERS (showing respect)
OBEDIENT (dutiful, meeting obligations)

Self-Direction Vglgg

FREEDOM (freedom of action and thought)
SELF-RESPECT (belief in one’s own worth)
CREATIVITY (uniqueness, imagination)
INDEPENDENT (self-reliant, self-sufficient)
CHOOSING OWN GOALS (selecting own purposes)
CURIOUS (interested in everything, exploring)

Pretest Attitude Toward Tuition Increase

I think that increasing tuition to hire more faculty is a good idea.

I am in favor of increasing tuition to hire more professors.

I believe that tuition should be increased to hire more professors.

I support raising tuition to hire more faculty.

The Board of Trustees should increase tuition to hire more professors.

Conformity Content
The types of students who support increasing tuition to hire more faculty.

The percentage of students who support increasing tuition to hire more
faculty.

The personality of students who support increasing tuition to hire more
faculty.

The agreeable nature of students who support increasing tuition to hire more
faculty.

42

PT]

.69

.55

.63
.47

.65

.68
.69

.70
.52
.71
.66
.52
.61

.90
.96
.96
.95
.93

.84
.75

.81

.76

Self-Direction Content
The benefits of increasing tuition to hire more faculty.

The freedom gained by students if tuition is increased to hire more faculty.
The abilities gained by students if tuition is increased to hire more faculty.

The independence gained by students if tuition is increased to hire more
faculty.

Message Qualifl

The information provided is believable.

The message is convincing.

I find the message to be compelling.

The information presented seems logical.

The conclusions drawn about raising tuition do not seem to be reasonable.
The reasoning used in the message was unsound.

Posttest Attitude

I think that increasing tuition to hire more faculty is a good idea.

I am in favor of increasing tuition to hire more faculty.

I believe that tuition should be increased to hire more faculty.

I support raising tuition to hire more faculty.

The Board of Trustees should increase tuition to hire more faculty.
I do not think tuition should be increased.

Qutcome Involvement
Whether or not tuition is increased is very important to my life at MSU.

The outcome of the proposal to increase tuition directly affects me.

The final decision regarding whether tuition increases or not will have an
impact on my life.

The outcome of the vote to increase tuition is not relevant to me.

Behavioral Intentions

I intend to vote in favor of the proposed tuition increase.

I will vote “Yes” for the proposed tuition increase.

Given the choice, I will vote to support the tuition increase.

I believe I will vote against the tuition increase.

If I were to vote today, I would vote against the tuition increase.
I plan to vote against the proposed tuition increase.

43

I’d

.87
.95
.89
.88

.82
.88
.71
.80

.57

.87
.96
.93
.97
.93
.74

.69
.91
.62

.65

.91
.94
.93
.90
.87
.88

APPENDIX C
STUDY 1 CORRELATIONS

Correlations between variables in path models for Study 1 (Conformity Message);
Corrected correlations are in the upper triangle

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

OD CONF SD MQ AMR BI
Other Directedness .13 .05 -.04 .29 .34
Conformity .09 .39 .26 .22 .26
Self-Direction .04 .29 -.10 -.01 -.03
Message Quality -.03 .20 -.08 .26 .30
Agreement with Msg Rec. .23 .18 -.01 .24 .95
Behavioral Intentions .27 .21 -.03 .27 .91
n = 72

Correlations between variables in path models for Study 1 (Self—Direction Message);
Corrected correlations are in the upper triangle

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

OD CONF SD MQ AMR BI

Other Directedness -.26 -.16 -.25 -.06 -.07
Conformity -. 18 .41 .06 -.21 -. 15
Self-Direction -. 12 .31 .28 .13 .01
Message Quality -.19 .05 .23 .66 .65
Agreement with Msg. Rec. -.05 -.17 .11 .60 .94
Behavioral Intentions -.06 -. 12 .01 .59 .90

n = 72

44

APPENDIX D

STUDY 2 MEASURES

Other-Directedness

In different situations and with different people, I often act like very different
persons.

I’m not always the person I appear to be

I may deceive people by being fiiendly when I really dislike them

Conformig Value
POLITENESS (courtesy, good manners)

SELF -DISCIPLIN E (self-restraint, resistance to temptation)
HONORING OF PARENTS AND ELDERS (showing respect)
OBEDIENT (dutiful, meeting obligations)

Self-Direction Val_ue_

FREEDOM (freedom of action and thought)
SELF-RESPECT (belief in one’s own worth)
CREATIVITY (uniqueness, imagination)
INDEPENDENT (self-reliant, self-sufficient)
CHOOSING OWN GOALS (selecting own purposes)
CURIOUS (interested in everything, exploring)

Pretest Attitude Toward Institutionalization

I think that institutionalizing people with severe mental illness is a good idea.
I am in favor of institutionalizing people with severe mental illness.

I believe that people with severe mental illnesses should be institutionalized.
I support institutionalizing people with severe mental illness.

Conformity Content
The agreeable nature of students who support institutionalizing people with

severe mental illnesses.

The personality of people who support institutionalization of people with
severe mental illnesses.

The types of students who support institutionalization of people with severe
mental illnesses.

The percentage of students who support institutionalization of the mentally
ill.

45

1'11

.62

.83
.50

.57
.58
.59
.64

.46
.54
.54
.58
.56
.58

.92
.97
.93
.98
.74
.85
.93

.79

Self-Direction Content

The benefits of institutionalizing people with severe mental illnesses.

The skills gained by people who are institutionalized with severe mental
illnesses.

The abilities gained by people who are institutionalized with severe mental
illnesses.

The independence gained by people who are institutionalized with severe
mental illnesses.

Message anlity

The message is convincing.

I find the message to be compelling.

The conclusions drawn about institutionalizing people do not seem to be
reasonable

The reasoning used in the message was unsound.

Posttest Attitude

I think that institutionalizing people with severe mental illness is a good idea.
I am in favor of institutionalizing people with severe mental illness.

I believe that people with severe mental illnesses should be institutionalized.
I support institutionalizing people with severe mental illness.

Psychiatrists should be able to institutionalize patients with severe mental
illnesses.

I do not think people with severe mental illnesses should be institutionalized.

Outcome Involvement

The final vote about institutionalizing people with severe mental illness is
pertinent to me.

Whether or not psychiatrists are allowed to institutionalize the mentally ill is
very important to my life.

The outcome of the vote to institutionalize mentally ill people directly affects
me.

The outcome of the vote to institutionalize mentally ill people would not be
relevant to me.

Behavioral Intentions

I intend to vote in favor of institutionalizing people with severe mental
illness.

I will vote “Yes” for the proposal to institutionalize people with severe
mental illness.

Given the choice, I will vote to support institutionalizing people with severe
mental illness.

I believe I will vote against institutionalizing people with severe mental
illness.

I plan to vote against institutionalizing people with severe mental illness.

46

lobe
MNI'Tr

.94

.88

.72
.53
.67
.75
.95
.93
.91

.93
.79

.50

.74

.77

.52

.93

.92

.93

.50

.65

Appendix E
STUDY 2 CORRELATION S

Correlations between variables in path models for Study 2 (Conformity Message);
Corrected correlations are in the upper triangle

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

OD CONF SD MQ AMR BI
Other Directedness .00 -.09 -.38 -.13 -.32
Conformity .00 .59 .03 .09 .05
Self-Direction -.06 .41 .26 .17 .28
Message Quality -.27 .02 .19 .62 .75
Agreement with Msg. Rec. -.10 .07 .14 .52 .87
Behavioral Intentions -.25 .04 .22 .61 .79
n = 72

Correlations between variables in path models for Study 2 (Self-Direction Message);
Corrected correlations are in the upper triangle

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

OD CONF SD MQ AMR BI
Other Directedness -.26 -.06 .35 -.01 .18
Conformity -.18 .57 -.29 .22 .05
Self-Direction -.04 .40 -.08 .22 .04
Message Quality .25 -.21 -.06 .53 .56
Agreement with Msg. Rec. -.01 .18 .18 .45 .83
Behavioral Intentions . 14 .04 .03 .46 .76
n=71

47

LIST OF REFERENCES
Bazzini, D., & Shaffer, D. (1995). Investigating the social-adjustive and value-
expressive functions of well-grounded attitudes: Implications for change and for

subsequent behavior. Motivation and Emotion 19, 279-305.

 

Chaiken, S. (1987). The heuristic model of persuasion. In M.P. Zanna, J .M.
Olson, & C.P. Herman (Eds.), Social influence: The mtario Smposium (V 01. 5, pp. 3-
39). Hillsdale, NJ: Erlbaum.

Chaiken, S., & Eagly, A. (1983). Communication modality as a determinant of
persuasion: The role of communicator salience. Journal of Personality and Social
Psychology, 45, 241-256.

DeBono, K. (1987). Investigating the social-adjustive and value-expressive
functions of attitudes: Implications for persuasion processes. Journal of Personality and
Social Psychology, 52, 279-287.

DeBono. K., & Hamish, R. (1988). Source expertise, source attractiveness, and
the processing of persuasive information: A functional approach. Journal of Personalig
and Social Psychology, 55, 541-546.

Dillard, J ., & Hunter, J. (1989). On the use and interpretation of the emotional
empathy scale, the self-consciousness scales, and the self-monitoring scale.

Communication Research 16 104-129.

 

Eagly, A., & Chaiken, S. ( 1993). The psychology of attitudes. Fort Worth: TX.:
Harcourt Brace & Company.

Fishbein, M., & Ajzen, I. (1975). Belief attitude intention and behavior: An

 

introduction to theory and research. Reading, MA: Addison-Wesley.

48

Fazio, R., Lenn, T., & Effrein, E. (1984). Spontaneous attitude formation. Social

Cognition, 2, 217-234.

Herek, G. (1986). The instrumentality of attitudes: Toward a neofunctional

theory. Journal of Social Issues 42 99-114.

 

Herek, G. (1987). Can functions be measured? A new perspective on the
functional approach to attitudes. Social Psychology Quarterly, 50, 285-303.

Johnson, 8., & Eagly, A. (1989). The effects of involvement on persuasion: A
meta-analysis. Psychological Bulletin, 106, 290-314.

Katz, D. (1960). The functional approach to the study of attitudes. Public

 

Opinion Quarterly, 24, 163-204.

LaFrance, B. (1998). The experiential functions of attitudes. Unpublished
doctoral dissertation. E. Lansing, MI: Department of Communication.

Lavine, H., & Snyder, M. ( 1996). Cognitive processing and the functional
matching effect in persuasion: The mediating role of subjective perceptions of message

quality. Journal of Experimental Social Psychology, 32, 580-604.

Lennox, R. (1988). The problem with self-monitoring: A two-sided scale with a

one-sided theory. Journal of Personality Assessment, 52, 58-73.

Maio, G., & Olson, J. (1994). Value-attitude-behaviour relations: The moderating
role of attitude functions. British Journal of Social Psychology, 33, 301-312.
Maio, G., & Olson, J. (1995). Relations between values, attitudes, and behavioral

intentions: The moderating role of attitude function. Journal of Experimental Social

Psychology, 31, 266-285.

49

Murray, 8., Haddock, G., & Zanna, M. (1996). On creating value-expressive
attitudes: An experimental approach. In C. Seligman & J. Olson (Eds) The psychology of
values: The Ontario smposium, Vol. 8 (pp. 107-133). Mahwah, NJ: Lawrence Erlbaum
Associates.

Petty, R., & Cacioppo, T. (1986). The elaboration likelihood model of persuasion.
In L. Berkowitz (Ed), Advances in experimental social psychology (Vol. 19, pp. 123-
205). San Diego, CA: Academic Press.

Petty, R., & Wegener, D. (1998). Matching versus mismatching attitude
functions: Implications for scrutiny of persuasive messages. Personalig and Social
Psychology Bulletin, 24, 227-240.

Rokeach, M. (1973). The nature of human values. New York: Free Press.

Rokeach, M. (1968). A theory of organization and change within value-attitude

 

systems. Journal of Social Issues 24, 13-33.

Sarnoff, 1., & Katz, D. (1954). The motivational bases of attitude change.
Journal of Abnormal and Social Psychology, 49, 115-124.

Schwarz, S. (1992). Universals in the content and structure of values: Theoretical
advances and empirical tests in 20 countries. In M. P. Zanna (Ed), Advances in

experimental and social psychology (Vol. 25, 1-65). San Diego, CA: Academic Press.

Scott, W. (1965). Values and organizations: A study of fraternities and sororities.

Chicago: Rand McNally.

Shavitt, S. (1989). Operationalizing functional theories of attitude. In A. R.
Pratkanis, S. J. Breckler, & A. G. Greenwald (Eds), Attitude Structure and Function (pp.

31 1-338). Hillsdale, NJ: Lawrence Erlbaum Associates.

50

Shavitt, S. (1990). The role of attitude objects in attitude functions. Journal of
Experimental Social Psychology, 26, 124-148.

Smith, M., Bruner, J ., & White, R. (1956). Qpinions and personality. New York:
John Wiley.

Snyder, M. (1974). Self-monitoring of expressive behavior. Journal of
Personality and Social Psychology, 30, 526-537.

Snyder, M., & DeBono, K. (1985). Appeals to image and claims about quality:
Understanding the psychology of advertising. Journal of Personality and Social
Psychology, 49, 586-597.

Snyder, M, & Gangestad, S. (1986). On the nature of self-monitoring: Matters of

assessment, matters of validity. Journal of Personality and Social Psychology, 51, 125-

139.

51

"‘111111111111111111