‘5”

9907

 

LIBRARY
Michiga. State
University

 

This is to certify that the
thesis entitled 9

Motivated Social Cognitions about Politics: The Effects of
Integrative Complexity and Involvement on Perceptions of
Political Messages

presented by
Hillary Cortney Shulman

has been accepted towards fulfillment
of the requirements for the

MA. degree in Communication

7/»; 6:;

Major Professor’s Signature
x403. 22L Zoo?

Date

MSU is an afiinnative-action, equal-opportunity employer

 

 

 

 

 

 

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/ClRC/DateDue.indd-p.1

 

MOTIVATED SOCIAL COGNTI‘IONS ABOUT POLITICS: THE EFFECTS OF
INTEGRATIVE COMPLEXITY AND INVOLVEMENT ON PERCEPTIONS OF
POLITICAL MESSAGES
By

Hillary Cortney Shulman

A THESIS
Submitted to
Michigan State University
in partial fulfilhnent of the requirements
for the degree of
MASTER OF ARTS
Department of Communication

2007

ABSTRACT
MOTIVATED SOCIAL COGNITIONS ABOUT POLITICS: THE EFFECTS OF
INTEGRATIVE COMPLEXITY AND INVOLVEMENT ON PERCEPTIONS OF
POLITICAL MESSAGES
By
Hillary Cortney Shulman
Studies of political involvement fail to yield consistent results with regard to

involvement’s effects on message perception. Possible reasons are the lack of an
explanatory cognitive mechanism that underlies involvement, and the failure to account
for what type of involvement is being elicited. The purpose of this study is to examine
how level of integrative complexity relates to different types of political involvement.
This research will then be applied to examine how this relationship effects perceptions of
contradiction present within political messages. It is thought that type of involvement
will explain the extent to which people bias the political messages they receive. Further,
integrative complexity may contribute to theories about who gets involved with politics
and the function this involvement serves. It was found that levels of integrative
complexity related to outcome-relevant and value-relevant involvement in a positive way.
It was also found that people with higher degrees of integrative complexity and outcome
or value relevant involvement perceived messages as less contradictory. This
relationship was augmented when the messages were from a favored politician. These

findings demonstrate that type, and degree, of involvement influence how people pay

attention to political messages.

ACKNOWLEDGMENTS

I think it is most important to first thank my advisor, Dr. Tim Levine, whose help,
insight, and numerous proofreading sessions made this project possible. His support, guidance,
and openness allowed me to gain the most out of this project and truly turn this paper into my
own. I would also like to thank my mother and father for listening to my ideas, and encouraging
and supporting me through every endeavor. I also feel the need to thank my extended family
here at Michigan State. Especially my roommate Jenny who stood by me throughout the
program and even helped me collect data when in a bind despite her busy schedule. Also to
Brandon and Lauren who first acclimated me so well to the University. Without you all my time
here would not have been as much fun.

I would also like to thank my committee, Drs. Chuck Atkin and Stan Kaplowitz, who
provided me with nothing but support and guidance. Without both of their attention to detail and
vast knowledge of subject matter, I would not have been able to get this project completed.
Finally, I would like to thank my undergraduate research assistants Alicia and Andrea for their

hard work, patience, and enthusiasm.

iii

TABLE OF CONTENTS

LIST OF TABLES ................................................................................... v
LITERATURE REVIEW ........................................................................... 1
College Students and Politics ............................................................... 4
Integrative Complexity ...................................................................... 5
Types of Political Involvement ........................................................... 10
The Value Pluralism Model ............................................................... 17
The Functional Approach to Attitudes ................................................... 20
Hypotheses .................................................................................. 23
METHOD ............................................................................................ 28
Instructions .................................................................................... 28
Procedure .................................................................................... 29
Measures .................................................................................... 3O
Integrative Complexity ........................................................... 30
Involvement Scale .............................................................. 31
Perceived Contradiction Between Messages .................................. 33
Opinion Change After Reconciling Information Presented ................. 33
RESULTS ............................................................................................ 34
Hypothesis One ............................................................................. 34
Hypothesis Two ............................................................................ 34
Research Question One .................................................................... 35
Research Question Two ................................................................... 35
Hypothesis Three ........................................................................... 36
Hypothesis Four ............................................................................ 36
Hypothesis Five ............................................................................ 37
Hypothesis Six .............................................................................. 38
Hypothesis Seven ........................................................................... 39
Research Question Three .................................................................. 41
DISCUSSION ....................................................................................... 43
APPENDICES

Appendix A: Integrative Complexity Coding Scheme
Appendix B: Involvement Scales

Appendix C: Questionnaire Items

Appendix D; Message Evaluation Questionnaire
Appendix E: Questionnaires by Issue

Appendix F: Messages Used in the Study

Appendix G: Tables

REFERENCES ...................................................................................... 80

iv

Table 1

Table 2

Table 3

Table 4

LIST OF TABLES

Correlations Between Integrative Complexity and Outcome-Relevant
Involvement by Political Party and Issue Type ................................ 7S

Correlations Between Integrative Complexity and Value-Relevant
Involvement by Political Party and Issue Type ................................ 76

Correlations Between Outcome-Relevant Involvement and Perceived
Message Discrepancy by Political Party and Issue Type ..................... 77

Correlations Between Value-Relevant Involvement and Perceived
Message Discrepancy by Political Party, Message Condition, and Issue
Type ................................................................................. 78

LITERATURE REVIEW

Are college students rational, irrational, or just apathetic when it comes to politics?
Researchers disagree about whether or not students are a demographic that merit campaign
attention (Eliasoph, 1998; Greenberg, 2003; Putnam, 1996). While it is easy to dismiss
young adults as being uncaring about political issues, historical events have repeatedly
proven inconsistent with this conclusion. Many student movements have had a significant
impact on the political climate of this country, and this shows that students have potential
impact But how can this potential be harnessed? This question is an interdisciplinary
concern, and many of the social sciences have attempted to address the roots of civic and
political engagement

One difficulty with this question, however, is that different approaches involve
different levels of analysis rendering integration difficult. For example, according to Social
Identity Theory (T aijfel, 1981; Turner, 1996; Turner, Hogg, Oakes, Reicher, & Wetherell,
1987), political involvement can best be understood as a group phenomenon that is driven by
self-categorization stemming from group membership (Cohen, 2003; Duncan, 2005; Buddy,
2001). Cognitive theorists argue that certain people are pre-disposed to political thought
(Burdein, Lodge, & Taber, 2006; Caprara, Barbaranelli, & Zimbardo, 1999; Dolan &
Holbrook, 2001; Greenng & Jonas, 1993; Golec, 2002; Lodge & Taber, 2005; Sidanius,
1978, 1988; Skitka & Tetlock, 1993; Tetlock, 1983, 1986; Tetlock, Peterson, & Lerner,
1996). Political scientists often focus on perceptions of legitimacy (Conover, 1988; Jennings,
1991; lost & Major, 2001), deprivation (see Relative Deprivation Theory, Davis 1959;

Crosby, 1976; Foster & Matheson, 1995), or oppression (Corning & Myers, 2002; Lee, 2002)

that incites political activism. An important limitation in each, however, is their failure to
amalgamate all levels of information processing from the individual level to the group level.

It is proposed here that both ability and motivation determine how and why college
students get involved in politics. Both individual-level and group level factors impact ability
and motivation. The specific focus here concerns how students make sense of contradicting
political messages and what factors account for these perceptual differences. The answer to
this question has important implications for the more general understanding of how people
process political messages and why people become politically involved.

One way to examine both motivation and ability is through the cognitive structures
that produce patterned information processing. Work on motivated social cognition has
demonstrated that certain people are motivated to think about issues in specific ways
(Battistich & Aronoff, 1985; Caprara et al., 1999; Greenberg & Jonas, 2003; Schaller, Boyd,
Yohannes, & O’Brien, 1995; Webster, 1993). One cognitive construct particularly relevant
to political thought is integrative complexity (Caprara et al., 1999; Tetlock, 1983, 1986;
Tetlock et al., 1996). Integrative complexity can be defined as the ability to, or degree to
which, a person is able to integrate different perspectives into a coherent and comprehensive
understanding (Golec, 2002; Tetlock, 1983, 1986; Tetlock et al., 1996; Rosenberg, 1988).
This is an ability in that certain people are more capable than others of achieving a more
integrated understanding of an event Integrative complexity is also a motivation because
this type of thought is cognitively effortful, and as such, people who tend to put in more
effort may naturally be more motivated to process political messages. This motivated social
cognition has been shown to be relevant in a political context because it motivates people to

scrutinize information contained in messages in distinct and potentially important ways.

Another factor affecting political motivation is involvement. Involvement has
generally been defined as the degree to which someone feels something is important
(Thomsen, Borgida, & Lavine, 1995). While strength of involvement is considered to be an
important factor in message perception, research has suggested that people are involved in
different ways (Johnson & Eagly, 1989, 1990), and that different types of involvement
translate into different motivations in message processing and information seeking behaviors
(Cho & Boster, 2005). Therefore, a second set of variables that will be explored is type and
strength of political involvement.

There are ample theories to draw upon when hypothesizing about how integrative
complexity and involvement relate to one another and to political message perceptions.
Because type of involvement has not been thoroughly explicated in a political context,
however, much of the research thus far is inconsistent Nevertheless, three theoretical
perspectives are likely to be especially useful. These perspectives include the Value
Pluralism Model (T etlock, 1986; Tetlock et al., 1996), the Functional Approach to Attitudes
(Katz, 1960), and Duel Processing Models (Heuristic Systematic Processing, Chaiken,
Liberman, & Eagly, 1989; Elaboration Likelihood Model, Petty & Cacioppo, 1986a, 1986b).

First, the idea of integrative complexity will be reviewed with a focus on showing
how all of these interdisciplinary ideas interrelate. Second, involvement type and strength
will be explored. Third, the Value Pluralism model will be discussed followed by the
Functional Approach to Attitudes. Tenets from the Duel Processing models will be
interspersed to provide for predictions regarding information processing. It is hoped that by
understanding the interplay between integrative complexity and involvement, light can be

shed about how college students perceive, and make sense of, political messages.

College Students and Politics

Intentionally, and conveniently, the participant pool of interest in this study is college
students. While most research in the social sciences use this demographic in order to
generalize, albeit with hesitance, to the general public, the current study is concerned
exclusively with this audience. This is because college students should be at an age when
sophisticated political thought can begin to be developed. Further, college campuses are an
environment conducive to the rapid spread of political views. Understanding how college
students can be motivated to think about politics, and understanding the causal mechanisms
that underlie their opinions can be of great value to future campaigns. The reason for this
logic follows.

Generally public opinion research does not give much credence to the political
opinions of college students. While no adequate explanation exists as to why this is the case,
both political parties appear to be relatively uninterested in this demographic’s political
opinion (Greenberg, 2003).

Thus, the current research will attempt to identify contextual idiosyncrasies that may drive
normative perceptions of politics or even motivate civic engagement for college students.
Empirical studies have formd that contextual idiosyncrasies such as biographical availability
(Corning & Myers, 2002; Lee, 2002), group membership through social networks (Conover,
1988; Coming & Myers, 2002; Huddy, 2001) and attitude specificity (Coming & Myers,
2002; Golec, 2003) significantly predict whether a person will engage in political behavior.
Therefore, questions tapping the above contextual factors are important in determining
whether an instigator of political involvement, specifically in college students, is attributable

to contextual and/or cognitive motivations. More specifically, it will be examined how

differences in involvement account for differences in political message perceptions. It is
important to understand how college students make sense of political messages, and how
patterns of interpretation can be explained through internal and external factors.

The importance of answering the above questions, and further explicating the
relationship between college students and politics lies in the tremendous influence this
demographic can potentially have on the political climate of the United States. According to
the US Census Bureau college students currently make up 9% of the country’s eligible voter
population (http://factfinder.census.gov/). Further, college students aged 18 to 24 years old
are anywhere from 20% to 30% more likely to vote than people of the same age not having

attended any college (httmflwwwcensus.gov/prod/2006pubs/920-556gb. With this

demographic information comes two very important implications. First, college students are

 

more capable of being mobilized to political action than the average person their age
(Greenberg, 2003). Second, 9% is a large enough percentage to be significantly influential
considering the margin of victory in the 2004 election was 2.46% (lm://elections.gr;r_u.ed_u_/).
Thus, the importance of this demographic must be realized Due to the existence of such
potential, it is surprising how little work has been done to truly understand the political
ideologies, normative perceptions, and information processing capacities in college students.
Integrative Complexity

The idea of studying integrative complexity in the political realm grew out of the idea
that to best understand how voters process political messages, their personalities, beliefs, and
values must be considered (Caprara et al., 1999; Tetlock, 1983, 1986; Tetlock et al., 1996).
One way to do this is by studying motivated social cognitions. The idea is that these

cognitions are influential because they guide information processing in patterned ways

(Battistich & Aronoff, 1985; Caprara et al., 1999; Greenberg & Jonas, 2003; Schaller, Boyd,
Yohannes, & O’Brien, 1995; Webster, 1993; Westen, Kilts, Blagov, Harenski, & Harnann,
(in press». Some motivated cognitions that have received a large amount of empirical
attention have been personal need for structure (Schaller et al., 1995), nwd for cognition
(Petty & Cacioppo, 1986a, 1986b; Schaller et al., 1995), and need for closure (Webster,
1993) just to name a few. While many of these have been examined in several different
contexts, integrative complexity as a motivated cognition fits in nicely as an influential, and
distinct, style that organizes one’s political identity. This is because one potentially crucial
component of political reasoning is the desire for a coherent, consistent, and well-integrated
political identity (Rosenberg, 1988; Tetlock, 1986; Tetlock et al., 1996). A review of the
relationship between integrative complexity and political ideology follows.

Researchers generally define integrative complexity as the degree to which an
individual can weigh all perspectives and then integrate ideas into a coherent position (Golec,
2002; Tetlock, 1986; Tetlock et al., 1996; Rosenberg, 1988; Van Hiel & Mervielde, 2003).
Integrative complexity varies as a function of three factors including individual
predispositions, situational factors, and internal states that temporarily affect information
processing (Golec, 2002; Jost & Major, 2001; Schroeder, Driver, & Streufert, 1967; Streufert
& Streufurt, 1978; Tetlock, 1986; Tetlock et al., 1996). This being the case, it is important to
recognize that integrative complexity is both an individual difference and an issue specific
variable that can be enhanced under motivation.

Rosenberg (1988) theorized that integrative complexity consists of the capacity to
decentrate and coordinate information. Decentration refers to the ability to focus on multiple

aspects of an object or the relationship between objects (Golec, 2002; Rosenberg, 1988);

whereas coordination is considered, “the ability to integrate different perspectives into a
coherent and comprehensive understanding of a given political situation” (Golec, 2002, p.
734). Thus, the complexity with which a person decentrates and coordinates information
determines the level of integrative political reasoning.

According to this conceptualization of political reasoning those possessing a high
level of integrative complexity are successful in their development of conceptual connections
among differentiated characteristics, and as a result are able to create a political
understanding that is both abstract and complex (i.e., systematic reasoning; Golec, 2002;
Rosenberg, 1988). On the other hand, people who rate lower in integrative complexity (i.e.,
linear reasoning; Rosenberg, 1988) tend to rely on simple evaluative rules to categorize
information and tend to be more susceptible to situational characteristics, which facilitate
more temporary and inconsistent viewpoints across issues (Tetlock, 1983, 1986; Tetlock et
al., 1996; Van Hiel & Mervielde, 2003). Further, they lack a defined perspective and are
only able to rely on simple generalizations and abstractions when fonning opinions (Golec,
2002; Rosenberg, 1988). Based on the aforementioned definitions, it is best to distinguish
levels of integrative complexity in terms of fundamental differences along three criteria as
defined in Golec (2002): (1) the degree of abstraction one can reach with regard to their
understandings of the world; (2) the ability to transcend a single perspective of political
reality and the ability to integrate a multitude of perspectives; (3) and the extremity and
generality in their understandings of norms, rules, and values.

While past research on integrative complexity has often operationally defined the
construct as high or low based on a median split, Rosenberg’s (1988) conceptualization

divides reasoning among five levels. These levels include, from lowest to highest:

sequential, advanced sequential, linear, advanced linear, and systematic political reasoning
(Rosenberg, 1988). In order to identify people as operating at one of these levels of
reasoning, a coding scheme was implemented that assessed how people answered questions
based on 16-dimensions. The purpose of these dimensions was to address how many causes
people attributed to a certain political event and the degree of interrelationship among causes.
It was reasoned that more causes and more integration was indicative of more advanced
reasoning. To test his hypotheses regarding the existence of five levels Of political thought,
Rosenberg (1988) brought people into the lab and conducted a series of interviews about
political and social issues of the participant’s choice. While the questions were initially
simple, their complexity increased throughout the course of the interview. The degree to
which participants could respond to complex questioning determined their level of integrative
complexity. Further, results from three empirical studies found that a participant’s structure
of thought remained constant through all interviews irrespective of the issue being addressed.
This demonstrates that Rosenberg’s (1 98 8) operationalization of integrative complexity
seems to reliably measure this construct as an individual difference.

The importance of recognizing different levels of reasoning is that political thought
and attitudes are the foundation of one’s political identity (Conover, 1988; Greenberg &
Jonas, 2003). An important goal in the development of this identity is the ability to establish
opinions that remain consistent across issues. Once these positions are established and
supported, a value system is developed (Katz, 1960) that canies implications for how one
will perceive and interpret future information that agrees or disagrees with a particular
attitude. Balance theories have demonstrated that when values and/or attitudes are threatened

by incongruent information an aversive state is created that people will actively seek to

alleviate (e.g., Cognitive Dissonance Theory; Festinger, 1957; e.g., Balance theory, Heider,
1958). As a result, pmposive thinking and complex cogrritions are evoked to negotiate
discrepancies to reinstate cognitive equilibrium (Katz, 1960; Festinger, 1957; Rokeach, 1960;
Skitka & Tetlock 1993). Further, research has demonstrated that in order to return to a state
of attitudinal homeostasis, a person will be motivated to engage in information processing
that is systematic or biased (Chaiken et al., 1989; Festinger, 1957; Petty & Cacioppo, 1986a,
1986b). Whichever route is chosen, it is important to recognize that with contradiction
comes reasoning that will attempt to reduce cognitive dissonance when confi'onted with a
disparaging message.

The idea that people are constantly dealing with, and negotiating, contradictions
rooted in political value conflict is the basis of the Value Pluralism Model (T etlock, 1986;
Tetlock et al., 1996). This model asserts that when dealing with conflicting values, the first
step people engage in is the creation of a value hierarchy that implicitly gives preferential
treatment to some values at the expense of others. This is the notion of a ‘value trade-oft"
(T etlock, 1986; Tetlock et al., 1996; Skitka & Tetlock, 1993). Depending on motivated
cognitions, however, the ease of creating this hierarchy may vary depending on an
individual’s integrative complexity. Some individuals may be able to expend more cognitive
effort to create a more integrated value system, which could have significant implications
concerning firndamental differences in the ways people with varying degrees of complexity
interpret political messages.

Based on the conceptual definition of integrative complexity, it is reasoned that those
who rate higher in integrative complexity will have the tendency to develop a more integrated

and coherent political ideology that is founded in logic that connects a multiplicity of political

issues and events. Those low in integrative complexity however, should not be motivated to
create such an organized and comprehensive political schema. Thus, political ideologies
should be structurally different as a function of integrative complexity. This difference is
under investigation in the current study because it is reasoned that understanding this
cognitive motivation can explain type Of political involvement that is most salient and,
equally important, the patterns of message processing as a result of this individual difference.
Types of Political Involvement

Being politically involved is not solely contingent on how complex a person is able to
think Involvement also needs to be considered. Involvement is generally defined as the
degree to which someone feels something is important (Thomson et al., 1995). Involvement
can be considered a collaborative function of both environmental and cognitive factors that
manifest themselves in the intensity of one’s attitudes, beliefs, and/or behaviors. Reasons for
involvement can be authentic interest (or cognitive needs), close ties to one’s self-concept, or
normative influence. Motivations for political involvement is an interesting realm with
which to study information processing because knowledge is so varied due to political
messages that are often ambiguous and contradictory across positions, politicians, and
campaigns.

Referring to previous assertions about levels of integrative complexity and structural
differences of political thought, the argument made here is that it becomes necessary to first
identify, and second to individually address, involvement motivated by complex cognition
versus involvement attributable to other factors. The supposition here is that a predominately
internal motivator for political identity isfundamentally different than ideologies motivated

by external factors, and information processing patterns will differ distinctly as a result To

10

follow will be an examination of how involvement type could be better explained and applied
to politics from a more cognitive perspective. By understanding the relationship between the
two, more accurate predictions can be made concerning information processing in favorable
and unfavorable messages.

A paradoxical relationship exists in the literature perhaps because of the failure to
account for cognitive motivations behind involvement type differences. Past research in
political science (Dolan & Holbrook, 2001; Rudolph, 2006), psychology (Chaiken et al.,
1989; Festinger, 1957; Johnson & Eagly, 1989, 1990; Petty & Cacioppo, 1986a, 1986b), and
communication (Cho, 2005; Cho & Boster, 2005) have looked at how high or low levels of
involvement affect information processing. An argument advanced by Johnson and Eagley
(1989, 1990), however, suggests that looking at involvement as a unidimensional construct
does not accurately portray how this construct operates. More recently, Cho and Boster
(2005) tested the reliability and validity of scales measuring involvement along the three
dimensions originally conceptualized by Johnson and Eagly (1989, 1990). Results indicate
that there are three different types of involvement, specified as: (1) value-relevant; (2)
outcome-relevant; and (3) impression-relevant involvement, and that these types all produce
different patterns of perceptions when combined with other variables (Cho & Boster, 2005).
Implications suggest that research in involvement should begin to distinguish between these
types of involvement in order to yield more explainable empirical findings. Another
implication is that if research could better understand the cognitions potentially underlying
different types of involvement, persuasive attempts could be more precise when targeting a

particular demographic.

11

The first type of involvement specified is termed value-relevant, and encompasses the
construct often known as ego involvement (Johnson & Eagly, 1989). Value-relevant
involvement can be conceptually defined as a psychological state developed by the activation
of attitudes that are the result of important personal values (Cho & Boster, 2005; Johnson &
Eagly, 1989, 1990). This involvement type is inextricably tied to one’s self-concept and as a
result has the potential to be threatened by contradictory information. Because the
motivation behind this type of involvement is to maintain and reinforce a positive self-image,
persuasion is often difficult when value-relevant involvement is high. Research has found
(Cho & Boster, 2005) that there is a positive association between value-relevant involvement
and attitude extremity. This is because in an effort to maintain a stable, and favorable, self-
concept it is likely that individuals will bias information in ways that consistently reinforce
their pro-existing schema. Further, this assertion about value-relevant involvement and
attitude extremity has some important ties to integrative complexity based on contradictory
findings currently existing in research. By examining the role of cogrritions on attitudes, the
need for a more detailed specification of involvement types may become clearer.

Van Hiel and Mervielde (2003) conducted a study investigating the relationship
between cognitive complexity and political extrenrism. The purpose of this study was to
reinvestigate the claim by Rokeach (1960) and Tetlock (1983), that high integrative
complexity was most indicative of moderate, and left-wing, political ideologies. Van Hiel
and Mervielde (2003) and Sidanius (1973, 1988) disagreed with these prior findings on two
accounts. They believe that left versus right wing ideologies should not be structurally
different from one another. Instead, they claimed that extremists should be higher in

integrative complexity for four main reasons: (1) more independence and less conformity

12

pressm‘es; (2) relatively high intelligence; (3) a larger information arsenal as a result of
increased involvement; and finally (4) a higher stress tolerance (Sidanius, 1978, 1988; Van
Hiel & Mervielde, 2003). Findings from Van Hiel and Mervielde (2003) as well as two
studies conducted by Sidanius (1978, 1988), supported the conclusion that political
extremism and integrative complexity were positively related.

How can research reconcile these discordant findings? First, it is important to
acknowledge that today, researchers tend to agree that strength of ideology is more
structurally similar than content in ideology. In other words, staunch Republicans and devote
Democrats are more structurally similar to one another than their moderate counterparts
despite an overt difference of opinion. As Tetlock (1986) states in his Value Pluralism
Model, what we think is not necessarily indicative of how we think. Thus, the remainder of
this paper will be making political distinctions in terms of high versus low political
involvement rather than democrat versus republican or liberal versus conservative.

The second area of debate is whether integratively complex reasoning leads to
extreme or moderate viewpoints. Given that previous research lacks specification about
involvement types, both viewpoints might be right. This is because neither perspective is
picking up on how cognitive structure leads to different types of high involvement More
specifically, being high in value relevant involvement can lead to extremist viewpoints
resulting from a heightened concern with self-concept, whereas rating high in outcome
relevant involvement (see below) may lead to more moderate ideologies marked by a
concern with what is actually correct The latter relationship is consistent with what Rokeach
(1960) and Tetlock (1983) were finding, in relation to integrative complexity. The former is

consistent with findings fiom Van Hiel and Mervielde (2003) and Sidanius (1978, 1988).

13

Unlike value-relevant involvement, Johnson and Eagly’s (1989, 1990)
conceptualization of outcome-relevant involvement is distinguished by a concern with the
accuracy (and consequently outcome) of one’ 5 opinion (Cho & Boster, 2005). Outcome-
relevant involvement can be defined as the extent to which an attitudinal issue is behaved to
have personal consequences for them (Cho & Boster, 2005; Johnson & Eagly, 1989, 1990).
It follows that if this involvement is dictated by the consequences of an attitude, it is more
likely that an individual will scrutinize a message in an effort to make the best decision
(Chaiken et al., 1989; Cho & Boster, 2005; Petty & Cacioppo, 1986a, 1986b). This
relationship is supported by both Dual Processing Models (Chaiken et al., 1989; Petty &
Cacioppo, 1986a, 1986b) which assert that if an issue is of high relevance to an individual,
they will be more likely to pay attention, seek out additional information, and to carefully
evaluate the message content. As a result, this type of information processing tends to be
more objective and less susceptible to situationally induced biases (Cho & Boster, 2005;
Webster, 1993). Thus, because finding an accurate outcome is more cognitively taxing than
biasing information to maintain a previous, and preferential, opinion, it is believed that
integrative complexity as a cognitive style complements outcome-relevant involvement

One recent study that lends support to the above assertion is Duncan’s (2005)
investigation on personal political salience. A personal political schema scale (PPS) was
created that measured an individual’s propensity to internalize historical political events, and
View these issues as being influential to one’s daily life (Duncan, 2005). This construct
relates to Johnson and Eagly’s (1989, 1990) conceptualization of outcome-relevant
involvement because it is reasoned that if someone believes political events impact their daily

lives, they should be more motivated to create an ideology that is beneficial but also accurate

14

in its predictive ability. Duncan’s (2005) PPS scale comprised of 25 historical political
events (e.g., Vietnam War, the Feminist Movement, 9/11, etc) in which the respondent was
asked to rate on a scale of one (not at all relevant) to three (extremely relevant) how
important certain events were to their daily lives. Results from this study indicate that that
high PPS scores were positively related to more developed positions across issues. Rather,
this integration of perspectives across issues manifested itself in a stronger attitude and
behavior link for high PPS scorers such that,

. . .high scorers on PPS may have reacted to the war attitudinally, emotionally, or

behaviorally. For example, when thinking about what kind of car to drive, high

scorers on PPS may consider firel efficiency as very important- not just for economic
reasons, but because of its connections to US. dependence on foreign-bought fuel

(Duncan, 2005, p. 974)

Aside fi'om the integrative link between war attitudes and type of car to buy, the PPS scale
also found a positive behavioral link between scores and tendency to “obsessively watch
CNN, engage in emotionally charged conversations about the war, or participate actively in
war protest or support activities” (Duncan, 2005, p. 974). Thus, the results of this study
imply that people who feel the need to integrate their political perspectives tend to rmderstand
the need for logical consistency between issues. A second implication that can be derived
from this study is that for people engaging in behaviors that correspond with an integrative
ideology, outcome-relevant involvement may be an important goal of information
processing-

The distinction between value-relevant involvement and outcome-relevant
involvement can be identified by the adherence to different cognitive goals, which manifests
in distinct differences in information processing. Rudolph (2006), states that motivated

reasoning is driven by either accuracy goals (i.e., outcome-relevant) or directional goals (i.e.,

15

value-relevant). He defines accuracy goals by the desire to reach an accurate conclusion, and
directional goals as the desire to reach a preferred conclusion (Rudolph, 2006). It should be
clear, based on these definitions, how analogous these concepts are to the first two types of
involvement explicated. The contribution of Rudolph’s (2006) research is that he applied
these ideas specifically to political attitudes by examining how the mentioning of political
parties affected attributions of responsibility. What this research found was that when
partisan ones were offered, biased processing occurred (i.e., directional goals) in the form of
responsibility attributions. Rudolph (2006) asserted that, “responsibility judgments lie at the
heart of representative democracy” (p. 99). This is a dangerous assertion because the
subjective interpretability of these judgments are susceptible to extreme bias. This is why
more empirical attention must be given to the source of these biases. Duncan (2005)
maintains that with involvement comes information seeking. Another purpose of the current
investigation will be to decipher what cognitive factors lead to highly selective information
seeking that appeases value-relevant and directional goals, and what factors contribute to
outcome-relevant, accuracy driven information processing.

One last area of consideration thought to motivate political involvement is
environmental influences and normative perceptions of public opinion. It was previously
discussed that the motivation to become involved with politics is not only a function of
internal factors, but can also be instigated by a highly politicized environment. Historically,
there have been incidences of political “hotbeds” that have arisen based on leadership that
efl‘ectively mobilized a specific population. Examples include the Free Speech Movement in
Berkeley, California and anti—war protests in Madison, Wisconsin, Washington, DC, and

Chicago, Illinois. These examples indicate seem to indicate that an important component of

16

political involvement is overt expression and behavior. As such, the last type of involvement
specified by Johnson and Eagly (1989, 1990) and empirically supported by Cho and Boster
(2005) is impression relevant involvement. This type of involvement is defined as being
concerned with one’s self-presentation (Cho & Boster, 2005; Johnson & Eagly, 1989, 1990;
Zimbardo, 1960). This concern is rooted in one’s basic need for approval, especially with
regard to the consequences of one’s actions. The primary motivator for this type of
involvement is cues present within one’s environment, and one’s perception of normative
behavior in a given context According to Zimbardo (1960), concern with the consequences
of communication about an issue can be just as significant in forming opinions as the other
two forms of involvement. An important implication of impression-relevant involvement is
that because the expression of opinions is rooted in the need for approval, position on issues
is susceptible to change based on environment. Thus, if college students perceive a
normative opinion regarding an issue, it is important to understand whether the opinion
individuals are arriving at is attributable to the facts of the issue or conformity to the social
standard.

The idea of communication appropriateness is further augmented in politics. Because
politics is fundamentally controversial, people are likely to be strategic about the contexts in
which they choose to disclose their political opinions. Due to this, it merits investigation
whether strength of political involvement, confidence in political knowledge, and perceptions
of deviance fi'om the social norm leads to voluntary disclosure of personal political opinions.
Understanding this relationship may account for student’s perceptions of theirs, and other
student’s political participation as well as accuracy motivations in political information

processing.

17

Thus far, motivated social cognitions and involvement type have been explicated. It
has been demonstrated how cognitive structures potentially provide a foundation for
understanding political information, and developing integrative reasoning about a multitude
of issues, events, and politicians. What needs to be further explained however, is how
cognitive structure and intensity, and type, of involvement foster the creation of value
systems. Through this explanation, the pervasiveness of idiosyncratic information processing
patterns will be discussed.

The Value Pluralism Model

The Value Pluralism Model (VPM) was advanced through the acknowledgement that
everyone holds beliefs and values that inherently conflict at some level of reasoning. Tetlock
(1986) and later Tetlock et al. (1996) posit that this truism is especially applicable to politics
such that policies that satisfy one value, often require sacrificing another. An example of this
might make this idea more clear. Take two values held sacred in American society, fieedom
and equality. These political buzzwords are deeply entrenched in the rhetoric that mobilizes
citizens to proclaim their patriotism and ostracize dissenters who become accused of not
holding these values in the highest esteem. Regardless of the fact many people openly accept
and embrace these values, realistically adherence to both is idealistic and almost impossible.
Education provides an apropos metaphor for this contradiction. Parents may assert that they
want equality for their children, however are quick to exercise that part of their fi'eedom
entails the choice of sending their kids to private schools, or paying absorbent taxes to enroll
them in the best public schools. How do people deal with contradictory values and
principles? The rest of this section will explicate what cognitive processes help justify these

types of policy decisions through tenets of the VPM.

18

The VPM assumes that underlying all ideologies are core or terminal values that
specify what the ultimate goal or consequence should be of message perception, policy
support, or ideological preference (Tetlock, 1986). This model is pluralistic in nature beeause
ideologies consist of multiple values that an individual must prioritize based on fiequency,
intensity, and/or relevance (T etlock, 1986; Tetlock et al., 1996). A value hierarchy is
imposed so that an individual can rely on a cognitive prescription that resolves the
discrepancy by prioritizing some values at the expense of others (Skitka & Tetlock, 1993;
Tetlock, 1986; Tetlock et al., 1996). Research has shown that people who are highly
involved with an issue have the tendency to implement greater integrative complexity when
faced with value conflict because they are more motivated to resolve complicated
contradictions (T etlock, 1983, 1986; Tetlock et al., 1996; Van Hiel & Mervielde, 2003).
While this assertion appears parsimonious, it neglects that message perception is a subjective
process that could be undertaken with different goals in mind. Recall Rudolph’s (2006)
assertions about motivated reasoning; there can be two outcomes of political information
processing, either the facilitation of accuracy or directional goals.

It is proposed that those who are high in outcome-relevant involvement should
scrutinize information in an attempt to arrive at the best possible conclusion. They will likely
be motivated to expend a great deal of cognitive effort, and to seek additional information
that will help them resolve an existing conflict In accordance with the original VPM,
individuals who are high in integrative complexity (i.e., ability), and are highly motivated by
outcome-relevant goals, should have more deliberate and developed value hierarchies than

people low in ability and/or motivation.

19

 

 

 

 

II...- t‘! Q

The revised VPM (Tetlock et al., 1996), however, is careful not to presume that
cognitive ability is sOlely responsible for the establishment of a value hierarchy. Rather,
Tetlock et al. (1996) assert that elements within a given context significantly contribute to the
level of attention paid to a specific discrepancy or issue. Specifically, the revised VPM
predicts that the only way a person will engage in integratively complex trade-off reasoning,
is if the following five conditions exist (Tetlock et al., 1996): (1) the two contradicting values
are of equal strength; (2) it is socially acceptable to consider the trade-off; (3) the person is
accountable in some way; (4) a decision is necessary; (5) and lastly, there are motives for an
accurate and a high quality decision. If these conditions are not met, Tetlock et al. (1996)
posit that an individual will disengage from the issue or use heuristic or peripheral processing
to make a quick, and likely biased, decision expending little cognitive effort

The guidance that the VPM offers is predicated on postulates from the Duel
Processing Models. It is important to recognize that information processing is a fimction of
motivation and ability. While intensity of political involvement comprises the motivation
component, ability is conceptualized by integrative complexity that manifests in a value
hierarchy that can be articulated and developed. Due to the existence of this hierarchy, and
the motivation to achieve an outcome-relevant ideology, information processing should be
guided by accuracy goals (Rudolph, 2006). In this vein, it is likely that participants take a
more pragmatic approach when interpreting discrepant messages, and should seek
information that reconciles the differences at hand, while being reasonably free fiom
selectivity and preferential bias. This way of organizing information is consistent with the
central or systematic route to information processing, marked by effortful and objective

message processing (Chaiken et al., 1989; Petty & Cacioppo, 1986a, 1986b). One way to

20

measure the occurrence of this type of processing is to look at participant’s evaluations of
contradictory messages from a favored, neutral, and opposing source. If rational processing
is occurring, the perceived discrepancy between messages should remain constant because
preferential biases are not effectively priming message evaluation.

While the VPM is an appropriate model for understanding how integratively complex
thinkers might create a political schema, it does not adequately address people with less
developed hierarchies. As a consequence, the VPM also fails to specify predictions about
what happens when emotional biases color one’s information processing in ways that are
preferential rather than accurate. Because political reasoning is not always objective and not
always motivated by accuracy goals, a different theoretic approach must be considered to
explicate the areas outside the scope of the VPM. Katz’s (1960) Functional Approach to
Attitudes provides a nice segue for understanding how affect leads to ftmdamentally different
patterns of information processing.

The Functional Approach to Attitudes

The VPM is predicated on the assumption that the sociopolitical environment,
consequences of a given issue, and the ability of the individual will together determine the
likelihood of whether a person will engage in systematic reasoning. In politics, however, all
of these conditions are rarely met due to the emotional valence of the topics discussed. A
persistent problem of public opinion and public policy is that it is difficult to convince the
average person to care about issues that are so far removed from their day-to-day life
(Eliasoph, 1998; Putnam, 1996). This being the case, however, there are still people who are
intensely involved in politics even though they are not political elites, nor dealing with the

consequences of public policy decisions. Katz’s (1960) Functional Approach to Attitudes

21

(FAA) asserts that attitudes serve different functions, and if one can accurately identify the
function, the mechanisms behind the attitude become much more clear. In this section, the
frmctional approach will be applied to the DPM in order to support predictions about
message perceptions in the current experiment.

According to Katz (1960), there are four main functions served by attitudes. The form
functions are: instrumental, ego-defensive, value-expressive, and knowledge. The
dimensions of these functions that produce attitude diversity and theoretically determine
behavior include: intensity, degree of specificity, degree of differentiation, and the strength
(or number) of linkages to a value system (Katz, 1960). While the FAA is a difl’erent model,
it is important to recognize how analogous much of the FAA’s components are to previous
conceptualizations of integrative complexity, involvement, and the VPM. The purpose of
explicating tenets of the FAA in the current study, however, is not to reaffirm and firrther
support the aforementioned relationships. Rather, this discussion of the FAA will be focused
on the ego-defensive function of attitudes, because this perspective is outside the scope of
previous models, but still highly relevant to the current experiment

An influential firnction of political ideology is self-identity (Cohen, 2003; Huddy,
2001; Rokeach 1960). It follows that when one’s identity is threatened, often in terms of
value clashes, it is likely that a person will respond in an ego-defensive way. Katz (1960)
defines ego-defense as a means through which an individual avoids facing issues that are
dangerous to their self-concept. Because this firnction is inextricably tied to one’s identity, it
has been shown that an individual will go to great lengths to remove threats and to protect
their ego. In the political realm, one’s ego is often threatened especially because common

campaign tactics include political slander and attack ads. Thus it follows that if a person is

22

highly involved or strongly affiliated with a political party or issue, they will undergo
different information processing techniques when the message supports, versus opposes, their
position of choice. Further, in order to make sense of negative information, a person may
bias their information processing in order to maintain cognitive consistency (F estinger, 1957;
Heider, 1958; Katz, 1960). Katz (1960) suggests that one way to deal with cognitive
dissonance, or to protect against its occurrence, is to submit to “irrational factors of distorted
perceptions and wishful thinking” (p. 166). This phenomenon is consistent to predictions in
the DPM asserting that if someone is highly involved in an issue, they are harder to persuade
(Chaiken et al., 1989; Petty & Cacioppo, 1986a, 1986b). This is because they are not as
willing to listen to and apply logical arguments that go against one’s preconceived ideology.
Ego-defensive information processing is markedly different from processing driven
by outcome-involvement or integrative complexity. While people driven by accmacy goals
and integrative complexity can be as hard to persuade as those ego-involved, the persuasive
difficulties arise fiom different reasons that political research has failed to quantify. Those
high in outcome-relevant involvement tend to know more information and, as a result, are not
as reliant on single sources of information present within a given message. Thus,
inconsistencies are often resolved through outside information rationally applied to the value
hierarchy, as opposed to blindly accepting the terms of the discrepant message. To reiterate,
this approach is indicative of systematic or central processing (Chaiken et al., 1989; Petty &
Cacioppo, 1986a, 1986b), according to the DPM. It is important to recognize that not all
forms of high involvement, or types of people, are privy or capable of such an analytic route.
Individuals who get involved in a way tied to their self-concept may not always have

the ability to rationally process information that challenges their belief system. The DPM

23

asserts that individuals must be involved and have the capacity, or ability, to correctly process
a given message (Chaiken et al., 1989; Petty & Cacioppo, 1986a, 1986b). If a person is
involved, but does not possess these prerequisite cognitive capacities, it is likely that a
heuristic or peripheral route to processing will occur (Chaiken et al., 1989; Petty & Cacioppo,
1986a, 1986b). Katz (1960) goes on to predict that when an ego-defensive function is
elicited, an individual will bias new information to fit into a pre-existing schema. The
implication of this is that peeple whose political identity and self-identity are synonymous are
more difficult to persuade because discordant information is threatening. Ifthreatening
information is compounded by the inability to rationally process information, DPM and FAA
predict that biased information processing will occur. Thus, people high in value-relevant
involvement will interpret a political message differently than people high in outcome-
relevant involvement

Throughout the course of this review the argument has been advanced that it is
important to examine the role cognitions play in the formation of coherent issue positions.
This knowledge facilitates the understanding of why a value system is formed, and how open
and resilient it may be to alternative information. Understanding how these constructs relate
to one another, and Operate, may shed some light on how people, specifically college
students, reconcile discrepant political messages. In an attempt to synthesize the above
information, the following section will specify the hypotheses guiding the current research.
Hypotheses

The current study will expose participants to a series of messages from a politician
that contradict one another. The purpose of this will be to understand how college students

make sense of contradictory political information messages. It will further be investigated

24

how cognitive factors predict who will get involved with politics, and how these two factors
motivate specific patterns in information processing.

The first set of hypotheses concerns the relationship between integrative complexity
and involvement type. Prior research is inconsistent in the relationship between these two
constructs; however it was argued here that this discrepancy is attributable to involvement
being measured as a unidimensional construct. If this is the case, the following two
hypotheses are theoretically supported by both Rokeach (1960) and Tetlock (1983) and by

Van Hiel and Mervielde (2003) and Sidanius (1978, 1988).

H1: Scores on integrative complexity are positively associated with ratings of

outcome-relevant involvement.

HZ: Scores on integrative complexity are negatively associated with ratings of value-

relevant involvement

Another theoretical shortcoming is the failure to understand impression-relevant
involvements role in message perception from a cognitive standpoint Because political
involvement is influenced by environmental factors, does impression-relevant involvement
change as a firnction of environment or can it be seen as stable? Are highly involved people
apt to be more or less self-conscious about their behaviors? What in the environment
potentially moderates this consciousness? Because answers to these questions remain to be
empirically determined, a second set of research questions will be advanced regarding
impression-relevant involvements relationship to motivated social cognitions and other

involvement types.

25

RQl: Is there a relationship between level of integrative complexity and impression-

relevant involvement?

RQ2: Is intensity of impression-relevant involvement related to intensity in value-

relevant and/or outcome-relevant involvement?

After an examination of the anticipated structural differences between involvement
types, the next step will be to understand how these differences influence message
perceptions. Specifically, this research is interested in understanding how people make sense
of discrepant political messages. Because so much political rhetoric is contradictory,
understanding how people view inconsistency has real-world utility. Generally, participants
will be asked to view a set of political messages fiom three sources. One source will be
democrat, one republican, and one politically neutral. The logic behind these three somces is
the presumption that the participant will view one as ‘favored’, one as ‘opposing’, and the '
neutral as indifferent. Each message will contain two statements made by a politician, on
different dates that contradict one another. The participant will be asked to rate the extent of
the contradiction, and will then be provided with information that adequately reconciles the
discrepancy. Because the act of reconciliation is firndamentally an integrative task, it is

thought that one’s level of integrative complexity should be influential such that:

H3: Participant’s scores on integrative complexity will be negatively associated with
the extent to which messages are perceived as contradictory.
H4: As scores on integrative complexity increase, participants will shift opinions less

in response to exposure to contradictory information.

26

The next set of hypotheses will specify how involvement type should theoretically
relate to discrepant message perceptions. Based on past research, it is believed that people
high in outcome-relevant involvement should attempt to reconcile messages in an accurate
way. People who are high in value-relevant involvement, however, should be motivated to
maintain cognitive consistency by viewing their preferred candidate more positively than the
other two. By being motivated by these directional goals, it is likely biased information
processing will occur and be distinguishable by differences in perceived discrepancy as a
firnction of political party. Further, preferential goals will hinder information processing such
that new positive information about an opposing source will be ignored, while new and
positive information about a favored source will be highly attended to. For students low in
involvement it is thought that discrepancy would be viewed as generally high, but will be
easily alleviated when reconciling information is presented. This is because of the lack of
scrutiny of message content, due to low motivation, and an over-reliance on message cues
due to the lack of outside information. It is unclear however, if environmental changes in
impression-relevant involvement can provide an adequate catalyst for motivation and

message scrutiny. Thus, the following hypotheses and research questions are advanced.

H53; Degree of outcome-relevant involvement should be negatively related to
perceived message discrepancy.
H5b: The degree of outcome-relevant involvement should also be negatively related

to degree of Opinion change after being presented with reconciling information.

27

H6a: Degree of value-relevant involvement should be negatively related to perceived
message discrepancy for the preferred source, and positively related to message
discrepancy for an opposing source.

H6b: The degree of value-relevant involvement will be negatively correlated with

degree of opinion change for preferred source.

H6c: The degree of value-relevant involvement will be positively related to degree of

opinion change for an opposing source.

H7: Participants low in all types of involvement should initially rate all messages as

highly discrepant, but should shift their position upon being presented with

reconciling information such that they will view the messages as being less
discrepant and the reconciling information as more accurate.

RQ3: Will high levels of impression-relevant involvement motivate accurate or

directional processing?

With the above hypotheses guiding the current research, it is important to reiterate the
purpose and importance of the current investigation. First, it is necessary to examine how
cognitive structures predict patterns in information processing. The second purpose is to
better understand how, and in what capacity, college students are involved in the political
process. The third is to examine whether traditional theories of political science, motivated
social cognitions, and attitudes are replicated when applied to an area as emotionally latent as
political message perception. Thus, insight into these facets of political psychology and
information processing can help research become better informed about the elusive,

inconsistent, and largely ignored yet highly sought after, college demographic.

28

METHOD
Participants

The sample of this study consisted of 122 undergraduate men (n = 69) and women (n
= 53). The mean age of the sample was between 19 and 20 years old (SD = .63).
Approximately 96% of the sample was from America, while 4% were international students.
Further 66% of the sample was White, 2% Hispanic, 6 % Afiican-American, 2% Asian, 1%
Native American, 1% Middle Eastern, and the rest of the sample considered themselves to be
other or their report was missing. Participants were enrolled in various classes at a large
Midwestem State University. All students received course credit for their participation in the
study.

In an effort to remain consistent with other procedures for measuring political
identity, self-report measures were included in this study. These included both categorical
and continuous measures of: political affiliation, strength of political affiliation, liberal versus
conservative attitudes and strength of attitudes, past voting behavior and intentions, and lastly
parent’s political opinions (See Appendix C). In terms of political participation, 53% of the
sample had voted in the past while 28% stated that they have not voted in the past. For
measures of political affiliation, 26% of the sample considered therrrselves Democrats, 16.7%
republican, 4% Independent, 3% Libertarian, 1% Green Party, and finally 30% stated they
had no political affiliations.

Instructions
Upon arrival to the laboratory, subjects were greeted by an experimenter, given a

brief overview to the study, and asked to fill out and sign a form for their voluntary consent

29

Participants were told that the purpose of the study was to help determine how college
students think about politics and political messages.
Procedure

The current study was a between subject 2 (Affirmative Action v. Capital
Punishment) by 2 (Republican v. Democrat) design, and consisted of three phases including:
writing about a political issue (either capital punishment or affirmative action), a structured
interview about that same issue, a survey gauging political involvement, affiliation, and
behavior, and lastly a series of messages from a political source.

During the first stage, participants were asked to write their thoughts about the
political issue assigned to them. Specifically, they wrote a paragraph about why the given
issue is debated, a list of both side’s arguments, and which political party is associated with
these arguments. After the paragraph was completed, the participant was interviewed.
Participants were brought into a room with an experimenter and told that their conversation
will be tape-recorded and that their identity will remain confidential. Before the tape-
recorder was turned on, the participant was told that they will be discussing the same topic
they just wrote about in the essay. The interviews lasted approximately five minutes, and the
interviewer read five questions fi'om a script and was not allowed to deliver any feedback
about parficipant’s responses to reduce social desirability issues.

The second phase of the experiment was a series of survey questionnaires gauging
political affiliations and behaviors, demographic characteristics, and involvement type scales
about the issue discussed and politics in general. The involvement type scale was a revised
form of Cho and Boster’s (2005) involvement scale that was altered to match the issues

specific to this study. This involvement scale was given to discriminate between outcome-

30

relevant involvement, value-relevant, and impression-relevant involvement There were two
indicators used to determine each involvement type. One indicator contained items
measuring involvement in politics in general, while the other included items measuring issue
specific involvement.

The third phase was exposure to a series of political messages. The participant was
told that the following three messages all came from either the head of the Democratic
National Committee (DNC) or the Republican National Comrrrittee (RN C). Assuming that
student’s identify with one of the major political parties (i.e. democrat or republican), it
follows that they perceived this message as being fi-om a favored source, or fi'om an opposing
source.

The participant was exposed to an initial statement made by the head of the DNC or
RNC. A contradictory statement then followed containing a message by the same target, that
contradicted the initial statement The participant was asked following message exposure a
series of questions examining the extent to which they perceived the manipulated
contradiction. The third page presented the participant with an exculpatory statement, which
used contextual information to explain the contradiction. As a second measure, participants
were asked to evaluate the contextual information. These procedures were the same across
conditions, issues, and message sources. After the participant was exposed to all four phases
of the study, they were debriefed.

Measures

Integrative Complexity. There were two methods used to obtain an individual’s
integrative complexity score. The first method asked the participant to write a paragraph

concerning their understanding about one of the political issues previously mentioned After

31

completion, the paragraph was scored in accordance with the integrative complexity coding
manual (Baker-Brown et al., 1992). The second method used to obtain complexity scores
only differed from the first in that the data analyzed was from the interview rather than the
written paragraph. Like the written part however, these data were scored using the
integrative complexity coding manual (Baker-Brown et al., 1992). The correlation between
scores for the written portion and interview portion was, r (1 19) = + .24, p < .01. Generally,
scores for integrative complexity range fi'om one being the lowest to seven indicating the
highest degree of complexity. For the current sample however, the range of scores was fi'orn
one to six for the composite measure of integrative complexity (M = 2.53, SD = .77), and
there was a slightly positive skew on the distribution.

In an effort to ensure that coding for integrative complexity was both valid and
reliable, two coders blind to hypotheses were trained in two stages. First, coders were given
a condensed version of the integrative complexity coding manual (Baker-Brown et al., 1992)
that also contained dimensions taken from Rosenberg’s (198 8) measmement of political
reasoning. Based on this information, coders were asked to read the materials and complete a
series of practice problems that were already coded by the authors of the manual. Once the
coders Independently proved that they were able to code reliably with the authors of the
manual (determined by the coding of problem sets as the same, or within one point of the
scores given by the authors; the procedure recommended by the manual), they were
considered to be validly measuring integrative complexity.

Once the coders were consistent with the manual, training sessions were conducted to
ensure that they were coding reliably with one another. Both coders scored all of the written

responses, and it was determined that the two coders were highly internally consistent, r

32

(120) = +85 and reliable, or = .92. After this, both coders each scored half of the interviews
Independently after scoring ten together to confirm reliability would be maintained when the
format shifted fiom written to oral response. A more detailed description of the numerical
values assigned to different levels of reasoning can be found in Appendix A.

Involvement Scale. In order to assess political involvement, Cho and Boster’s (2005)
three dimension involvement scale was used. Cho and Boster’s (2005) original measurement
assessed outcome-relevant, value-relevant, and impression-relevant involvement by asking
subjects to rate levels of importance assigned to specific issues (e. g., marijuana legalization,
abortion, death penalty, toothpaste, and jeans). In the current experiment, these questions
were replicated using one of the two issues previously mentioned (Appendix E), and were
followed by a general political involvement scale (See Appendix B). Both scales were
analyzed Independently for reliability, and then later averaged to obtain an involvement type
composite score.

There were approximately nine items gauging outcome-relevant involvement in the
original scale (Cho & Boster, 2005). This scale included questions such as, “It is easy for
me to think of ways that political issues affect my life,” and “My life changes are a direct
result of capital punishment legislation.” Questions such as this attempted to adequately
measure whether the participant felt that political involvement and issue involvement were
both personally relevant. In the issue related outcome-relevant involvement scale, all nine
items were kept with strong reliability, or = .86. The general political outcome-involvement
scale yielded a reliability of or= .87 when one item was omitted fiom measurement Results
obtained for the composite measure of outcome- relevant involvement revealed a normal

distribution for this measure (M = 3.91 , SD = 0.98).

33

This measurement also included nine items measuring value-relevant involvement.
Sample items included, “My political position is based on the values with which I try to
conduct my life,” and, “My position on Affirmative-Action reflect who I am.” Again, two
versions of this scale were used, one measuring issue specific value-relevant involvement (or
= .89) and the other measuring general political involvement (or = .87). All items were kept
in both of the scales. Results obtained for the composite measure of value-relevant
involvement showed a slightly normally skewed distribution (M = 4.36, SD = 1.15).

Finally, in Cho and Boste'r’s (2005) study five items were used to assess impression
relevant involvement Examples from this dimension include, “People may judge me on the
basis of the opinion that I express in public about political issues,” and, “If I express the right
kind of opinion about Affirmative-Action, people will find me more attractive”. For issue-
specific impression relevant involvement, two items were omitted leaving three to be scaled
(or = .81). Only one item was orrritted that gauged general political impression relevant
involvement (or = .80). The composite measure of impression relevant involvement revealed
a normal distribution (M = 4.28, SD = 1.19).

Perceived contradiction between messages. One of the dependent variables was the
amount of discrepancy the participant attributes to the message source (See Appendix D).
The participant was asked a series of questions regarding their perceptions of the
contradiction such as, “To what degree do you feel that the previous two messages
contradicted each other?” and was given a seven-point Likert scale ranging from 1 (strongly
contradict) to 7 (not at all contradictory). It was thought that the favorability of the message
source would weigh into the participant’s willingness or motivation to reconcile the

discordant viewpoints.

34

Opinion Change afler reconciling information presented A second measure of
perceived discrepancy was used to better understand how the participant processes new
information. After the participant rated the perceived level of inconsistency between
messages, they were provided with information that reconciled the discrepancy and explained
why the source changed their mind. After exposure to this information, the participant was
asked a series of questions gauging their perceptions of the reconciling information such as,
“I think the new information accurately explains how the first two messages relate”. Again a

Likert scale, with the same anchors, was used to quantify responses.

35

RESULTS
Hypothesis One

It was hypothesized that scores on integrative complexity would be positively
associated with ratings of outcome-relevant involvement. The correlation between
integrative complexity and outcome-releVant involvement was significant, r (120) = +20, p <
.05. Thus the findings were consistent with predictions for hypothesis one.

An unpredicted finding, however, was that results differed for participants who
reported themselves as being republican versus participants who did not affiliate with any
political party. Although the correlation between integrative complexity and self-reported
Republicans was not statistically significant, r (23) = -.16, p = .442, the reported correlation
between integrative complexity and those not affiliated with any political party was
significant, r (44) = + .43, p < .05. Further, these two correlations were significantly difl‘erent
fi'om one another, 2 = 2.38, p < .05. A more detailed description of this data can be found in
Table 1 located in Appendix G.

Hypothesis Two

The second hypothesis posited that integrative complexity scores would be negatively
related to ratings of value-relevant involvement Results found that the correlation between
integrative complexity and value-relevant involvement was significant ( r [120] = +. 19, p <
.05), in the direction opposite of the hypothesis. Thus, findings were inconsistent with initial
predictions, suggesting that participants higher in integrative complexity were also likely to
score higher in value-relevant involvement.

It should be noted that correlations between integrative complexity and value-relevant

involvement significantly differed based on issue, 2 = 2.28, p < .05. The correlation found

36

between integrative complexity and value-relevant involvement in the affirmative action
conditiOn was r (59) = +35, p < .05, whereas in the capital punishment condition the
correlation between the same two constructs was, r (59) = -.06, p > .05. The difference
between these two correlations by issue is largely attributable to significant differences
between participants reporting no political affiliation. For the affirrnative action condition,
those reporting no political affiliation yielded a significant and large integrative complexity
and value-relevant involvement correlation, r (15) = +60, p < .05. This finding significantly
differed from the capital punishment condition, 2 = 2.65, p < .05. However, for those with no
affiliatiorr, the correlation between integrative complexity and value-relevant involvement in
the capital punishment condition was not significant, r (27) = -.19, p = .337. These findings
are more specifically depicted in Table 2 in Appendix G.
Research Question One

The first research question asked whether there would be a relationship between
integrative complexity and impression-relevant involvement. A correlation between the two
revealed that there was no significant relationship between the two variables, r (120) #09, p
= .312. This finding suggests that level of integrative complexity is not clearly associated
with reported differences in impression-relevant involvement.
Research Question Two

The second research question asked whether impression-relevant involvement related
to outcome-relevant involvement and/or value-relevant involvement. Correlations revealed
that the relationship between impression-relevant involvement and outcome relevant
involvement was significantly positively correlated, r (120) = +.45, p < .001. Furthermore,

there was a significantly positive relationship yielded between impression-relevant

37

involvement and value-relevant involvement, r (120) = +50, p < .001. Thus, findings fipm
this study suggest that as impression-relevant involvement increases, so should scores in
outcome-relevant and value-relevant involvement.

Hypothesis Three

The third hypothesis asserted that there would be a negative correlation between
integrative complexity scores and the extent to which participants perceived the initial two
messages as being contradictory. Using each scale item as a dependent variable due to
scaling difficulty, one significant result was formd and was consistent with the initial
hypothesis. There was a significant negative correlation, r (122) = -.22, p < .05, between
integrative complexity scores and the fourth item that asked participants to what extent do the
two messages, “have nothing to do with one another”. Thus, the more highly participants
scored on integrative complexity, the less likely they were to think that the two messages
were unrelated. This finding was consistent with this hypothesis.

Upon looking at the results more closely, there are a few other significant
relationships to consider with regard to this hypothesis. It was found that for the issue of
affirmative action, there was a significant correlation between integrative complexity and
item number three, which asked participants to what extent were the two statements, “not at
all contradictory”, r (59) = -.25, p < .05. Thus, consistent with the aforementioned finding,
the greater one’s integrative complexity, the less likely one is to View the two messages as
contradictory.

Hypothesis Four
It was hypothesized that as scores on integrative complexity increased, participants

would shift their opinions less in response to reconciling information. A correlation revealed

38

that this hypothesis was not consistent with findings, r (120) = +.16, p = .089. Thus,
hypothesis four was not supported because it appears that level of integrative complexity
does not clearly relate to opinion change.

Hypothesis Five

Hypothesis 5a posited that outcome-relevant involvement was negatively related to
perceived message discrepancy. When analyzing this data, it was found that the correlation
between outcome-relevant involvement and perceived message discrepancy was not
significant when interpreted using both issues in aggregate, r (120) = -.06, p = .497.
However, when splitting the data up by issue, it was found that a significant negative
correlation existed between outcome-relevant involvement and message contradiction in the
case of capital punishment, r (5 9) = -.29, p < .05. Therefore, findings for capital punishment
were consistent with the initial hypothesis such that as outcome-relevant involvement
increased, perceptions of message discrepancy decreased. These findings are depicted in
Table 3 in Appendix G.

Furthermore, data suggests that political affiliations may have also affected the
relationship between outcome-relevant involvement and perceptions of message discrepancy
in the case of capital punishment. The strongest association existed between the above-
mentioned constructs in the case of reported Democrats, r (15) = -.74, p < .01. Although this
correlation did not significantly differ fi'om Republicans and ‘other’ political parties, it was
found to be statistically different from participants reporting no political affiliations, z = 2.64,
p < .01. For participants reporting no political affiliations, almost no association was found
between outcome-relevant involvement and perceptions of discrepancy, r (27) =+ .08, p =

.668. These findings suggest that people who affiliate with political parties perceived the

39

messages as more discrepant based on their outcome-relevant involvement than participants
reporting no such affiliations.

Hypothesis 5b suggested that degree of outcome-relevant involvement should also be
negatively related to degree of opinion change after the presentation of reconciling
information. Findings however failed tolbe consistent with this hypothesis, r (120) = -.04, p
= .673, which suggests that degree of outcome-relevant involvement does not affect
participants perceptions, or acceptance, of the reconciling information.

Hypothesis Six

In the first part of hypothesis 6a it was predicted that value-relevant involvement
should be negatively related to perceived message discrepancy for a preferred source (i.e.,
same political afiiliation as participant). A correlation revealed that the relationship between
value-relevant involvement and perceptions of message discrepancy was significant in the
direction predicted, r (29) = -.36, p < .05. Thus, as value-relevant involvement increases,
perception of message discrepancy decreases when the source of the message is a politician
from one’s preferred affiliation.

The second part of hypothesis 6a speculated that value-relevant involvement should
be positively associated to message discrepancy for an opposing source (i.e., a politician fi'om
an opposing political party). Results found that when both issues were analyzed together,
findings for this hypothesis were not consistent with initial predictions. When issues were
separately analyzed however, it was found that in the case of capital punishment, there was a
significant negative correlation between value-relevant involvement and perceptions of
message discrepancy, r (12) = -.65, p < .05. Thus, results for this issue revealed a significant

relationship in the direction opposite of the one predicted. Results for affirmative action,

40

however, approached significance in the direction originally hypothesized and opposite from
capital punishment, r (17) = +.4 1 , p = .084. Therefore, data suggests that both the issue and
the degree of value-relevant involvement affect how people interpret messages fi'om an
opposing candidate. A more detailed description of how perceptions of contradiction varied
by political affiliation can be found in Table 4.

Hypothesis 6b asserted that value-relevant involvement should be negatively
correlated with degree of opinion change when the message was flour a preferred source.
The results indicate that although the relationship between value-relevant involvement and
degree of opinion change were in the predicted direction, the correlation only approached
statistical significance, r (29) = -.33, p = .073. Thus, according to this study, no conclusion
can be made as to whether a relationship exists between value-relevant involvement and
opinion change when the source of the message is a preferred politician.

The predicted correlation in hypothesis 6c specified that value-relevant involvement
should be positively related to degree of opinion change when the message was fiom an
opposing source. Results revealed that the relationship between value-relevant involvement
and opinion change was not significant, r (31) = +.02, p = .928. Further, when examining
whether the correlations differed between value-relevant involvement and opinion change for
a favored versus an opposing source, results came out insignificant Thus, these findings
suggest that the relationship between value-relevant involvement and opinion change do not
appear to be affected by whether the source of the information is fiom a favored or opposing

politician.

41

Hypothesis Seven

The first part of hypothesis 7 predicted that participants low in all types of
involvement would initially rate all messages as being highly discrepant. In order to test this
hypothesis, two statistical analyses were conducted. In order to conduct the first, participants
were split into two groups indicating whether they were generally high in political
involvement or low. The question tested was whether those who were low in all types of
involvement rated the messages as being more discrepant than participants rating high in
involvement. Participants were labeled as low in involvement if the product of all three of
their involvement averages came out to less than 64. The score of 64 was chosen because
this is the product of the midpoints (i.e., four) of all involvement scales (i.e., 4 x 4 x 4). A
participant was labeled high in involvement if the product of their involvement scores were
64 or higher. After labeling participants as high or low, an Independent t-test was used to test
the hypothesis. When all cases were considered, it was found that participants low in
involvement indicated that the messages were significantly more discrepant (M = 5.68, SD =
1.32) than participants high in involvement (M= 5.22, SD = 1.64, t [120] = 1.67, p < .05,

772 = .02 ), using one-tail to set significance.

When splitting the data by issue however, it was found that when looking at the issue
of capital punishment there were significant differences between reported discrepancy based
on level of involvement, t (59) = 2.62, p < .05, 772 = .10. Specifically it was found that
participants low in involvement found the messages to be more discrepant (M = 6.01 , SD =
1.20) than participants considered high in involvement (M = 5.00, SD = 1.82). This

relationship was especially true for self-labeled Democrats in this sample regardless of

42

political issue. Results fi'om the study show that Democrats low in involvement perceived
both messages as being significantly more discrepant (M = 6.02, SD = .85) than Democrats
high in involvement (M= 5.17, SD = 1.66, t [32.6] = 2.08, p < .05, 77’ = .12. These two
findings are consistent with first part of hypothesis 7.

Part two of hypothesis 7 asserted that people low in involvement would show greater
opinion change upon being presented with reconciling information. This opinion shift would
be evident in participants low in involvement reporting less message discrepancy in time two
than participants high in involvement. First, in order to test this prediction, an Independent-
sample t-test was used to determine whether degree of opinion change was different based on
level of involvement. Based on this test, it was found that there was no significant mean
difference between level of involvement and degree of opinion change, I (120) = 0.15, p =
.885. Thus, the second part of hypothesis 7 was inconsistent with results from this study.

The final part of hypothesis 7 asserted that people low in involvement would view the
reconciling information as more accurate than people high in involvement. Again an
Independent-samples t-test was used. The result did not yield a significant difference
between high and low level involvement, t (120) = 0.16,p = .870. Participants low in
involvement did not find the new information to be more accurate (M = 4.82, SD = 1.23) than
participants high in involvement (M = 4.78, SD = 1.22). Based on findings, it can be
concluded that although level involvement did appear to affect participants initial perceptions
of the messages in the direction predicted, it cannot be concluded that level of involvement

affects opinion change and perceptions of accuracy after the presentation of new information.

43

Research Question Three

The flrird research question inquired whether high levels of impression-relevant
involvement would motivate accurate or directional information processing. In order to
examine this question, cases were selected for analysis if participants scored higher than four
in impression-relevant involvement. All other cases were excluded from analysis. To
explore whether processing was accurate or directional, an Independent-samples t-test was
used to compare mean scores between favored and opposing party in terms of perceived
initial contradiction, opinion change, and perceptions of information accuracy. The analysis
revealed that participants high in impression-relevant involvement did not significantly report
different scores for initial message discrepancy (t [3 8] = 0.65, p = .518) based on whether
they were reading a message from a favored (M = 5.02, SD = 1.72) versus an opposing
source (M = 5.38, SD = 1.66). Furthermore, it was found that participants did not shift their
Opinion any differently (I [38] = 0.55, p = .584) based on whether the source was a favored
(M= .11, SD = 1.01) or an opposing politician (M= .31, SD = 1.18). The only difference that
approached significance was information accuracy, I (38) = 1.87, p = .069. It was found that
participants high in impression-relevant involvement perceived opposing party information
as being slightly more accurate (M = 4.92, SD = 1.01) than information fi'om a favored party
(M = 4.25, SD = 1.24). These findings insinuate that participants high in impression-relevant
involvement perceive political information in more accurate, rather than directional, ways.
This assertion is evidenced by this data because participants high in this type of involvement

do not appear to process information any differently based on their political affiliations.

44

DISCUSSION

The purpose of this investigation was to understand how selected factors affect an
individual’s political attitudes as well as their perception of political messages. It was
thought that the motivated social cognition related to integrative complexity would affect a
person’s tendency to become involved in politics. Because the construct of integrative
complexity is thought to be related to one’s degree of political sophistication, it was thought
that people who scored higher in this reasoning ability would perceive political messages
differently than those scoring lower. Additionally, if integrative complexity shapes how a
person thinks about an issue, it also merited attention to investigate the relationship between
complexity and involvement type when looking at a person’s political attitudes.

The data fiom this study suggest that integrative complexity and type and strength of
involvement do affect how participants interpret political information. It was reasoned that
both involvement and integrative complexity would motivate a person to scrutinize political
information in an attempt to refine their political opinion The reason why understanding this
assertion is important is because people who are motivated to achieve a comprehensive
political opinion will reason through political messages differently than people who consider
themselves more apolitical.

Further, these findings demonstrate that an important factor in the political process is,
not surprisingly, the source of the political messages and the political affiliations of the
participant. It was found that people’s motivation to scrutinize the information they received
varied as a fimction of message source. When participants’ own opinions were threatened,
based on their favored candidate appearing contradictory, the response to the message was

quantifiably different than when the message source was the opposition. Specifically,

45

participants were able to reason through contradictions presented by their favored candidate
and reported these messages as “less contradictory” than the same set of messages presented
by an opposing candidate.

With so muCh media attention on campaigns and the character of the politicians
involved, an important question raised is how do people make sense of discrepant, or
inconsistent, messages fiom both favored and opposing political parties. For example, past
campaigns, such as the 2004 Bush strategy, have used a ‘flip-flop’ platform in an attempt to
undermine the credibility of the opposing candidate. The secondary purpose of this study
was to understand how political messages were viewed when contradiction did occur.

In order to test these empirical questions, first integrative complexity was measured.
Participants were asked to begin reasoning about a political issue by writing down the
arguments for and against that issue. Participants were also asked to identify which political
party affiliates with specific arguments. Secondly, participants were interviewed so that they
could elaborate on their understanding of the issue. Multiple indicators for the construct of
integrative complexity were important here because this ensured that all participants, whether
stronger writers or stronger speakers, would have the opporttmity to present their opinions
free fi'om method-specific constraints.

It is recommended that firture examinations into the construct of integrative
complexity explore how slight methodological differences such as interviews or writing
effect how this construct is inevitably measured. The current study found that, although
similar, the two different mediums conesponded with differences in coding. Specifically, it
was found that participants scored higher in integrative complexity in the interview portion of

the study than in the written portion. However, in the written portion one person was able to

46

achieve a score of six, while in the interview portion five was the highest score given.
Interpretations of these findings potentially hold implications about the construct validity of
integrative complexity measurement. One explanation of the data is that participants who
took more time on the written portion may have been more involved with the topic, which
increased their complexity score. In the interview portion, participants may have felt
pressure to hurry their response due to the presence of an interviewer and this pressure could
have hindered the quality of their response. On the other hand it could be argued that
impromptu responses best illustrate integrative complexity, indicated by the slightly higher
average score. Regardless, this study illustrates the controversy of using different methods to
obtain integrative complexity and in doing so poses the question of which method is the more
valid indicator of this construct

The third aspect of the study was a questionnaire that gauged involvement type for
both politics in general and the issue in particular. Again, the two indicators used to measure
involvement yielded slightly different scores. It is urged that future consideration be given
into what the wording on questionnaires implies for politics and public opinion. Results fi'om
this study found differences based on reports about the issue versus politics in general, and
this was important to consider before making assumptions about the type of political
participation an individual is reporting.

The final aspect of the study showed two messages that were identical fi'om both the
republican and democrat party. The two messages contradicted one another about the
politician’s perspective on the issue, and the reader of these messages was asked to what
degree do they perceive these statements as contradictory. Shortly thereafter contextual

information was given to the participant that would allow them to make sense of the initial

47

contradiction if they so chose. The participant was then asked whether this potentially
reconciling information: explained the contradiction, was accurate, was related to the issue,
or was biased in any way.

The first two hypotheses predicted how integrative complexity would be related to
outcome-relevant involvement and value-relevant involvement. For hypothesis one it was
found that participants scoring higher in integrative complexity also scored higher in
outcome-relevant involvement. Outcome-relevant involvement is thought to be involvement
motivated by accurate processing and the need to make ‘right’ choices. Thus, it is normally
associated with less biases and more openness to opinion change. These characteristics make
sense when understanding what motivates integrative complexity is the need for a well-
orgarrized and patterned way with which to view political events. Further, because
integrative complexity measures how well participants are able to integrate different aspects
of a political event into a coherent meta-understanding of the issue in general, it seems logical
that outcome-relevant involvement would facilitate this cognitive goal.

Another aspect considered when looking at integrative complexity was an attempt to
resolve debates in the literature regarding the affiliations of those higher in this construct
Results from this study found that Republicans actually showed a non-sigrrificant negative
correlation between integrative complexity and outcome-relevant involvement, while people
with no reported political affiliation showed a significant and substantial positive correlation.
This suggests that party affiliations may present a barrier between reasonable and accurate
information processing. This finding is explainable because if people want their preferred

party to be ‘correct’ (a processing technique indicative of value-relevant involvement), they

48

may dismiss evidence to the contrary and this pattern of information processing is not
indicative of sophisticated reasoning.

The second hypothesis examined the relationship between integrative complexity and
value-relevant involvement. It was predicted that there would be a negative relationship
between these two constructs because value-relevant involvement is more closely tied to the
ego-defense functions served by holding particular attitudes. Thus, inherent in this
involvement should be dedication to party affiliations and the need to view preferred parties
as better than opposing parties. The need to maintain preferential consistency seemed to be
contradictory to information processing patterns motivated by integrative complexity.
Results from the study, however, yielded a positive correlation between integrative
complexity and value-relevant involvement, which contradicted the argument presented in
the second hypothesis. Although not as strongly correlated as outcome-relevant involvement,
this finding suggest that people who are reasonably patterned and sophisticated in their
political thought still show a degree of ego invested involvement with their political attitudes.
Following are two post-hoe reasons why this finding showed up as significant in the opposite
direction.

A first potential explanation is that Cho and Boster (2005) found that outcome-
relevant involvement and value-relevant involvement are significantly positively correlated,
although the strength of association tended to vary by issue. This finding was replicated in
this current study. Thus, statistically if such a high positive correlation exists between these
two types of involvement, it is almost inevitable that a positive correlation would exist
between both constructs, or neither construct, in the case of involvement type and integrative

complexity. Thus, the relationship found in the present study might be spurious.

49

Second, this relationship appears logical when thinking about how and why people
hold their political viewpoints. Because politics is value-laden, it seems that teasing apart
objective versus subjective reasoning could be almost impossible. In other words, a negative
relationship between these two types of involvement would be idealistic because people who
believe they hold the most correct political opinion also likely believe, or uphold, the values
underlying those opinions. Because political attitudes often reflect the person holding the
attitude, teasing apart accurate or directional processing may be a falsely dichotomous way of
rmderstanding how involvement type effects political processing.

The first two research questions inquired about the relationship of impression-
relevant involvement and integrative complexity, as well as impression-relevant
involvements relationship to the other two forms of involvement Unlike outcome-relevant
or value-relevant involvement, impression-relevant involvement is rooted in an individual’s
concern with how their political attitudes are perceived by other people. Important to this
type is whether a person is intentionally conforming or intentionally disassociating fi'om
another group with known political attitudes. Because impression-relevant involvement is
concerned with appearance more so than political conviction, the relationship between this
involvement and patterns of reasoning could not be persuasively predicted, and was thus left
as a research question. This study found that the reason why this relationship could not be
easily predicted was, perhaps because no clear relationship exists between impression-
relevant involvement and integrative complexity. This suggests that how someone wants
their political opinions to be perceived does not hold implications for the sophistication of

their political attitudes.

50

The second research question exploring the correlation between impression relevant
involvement and the other two types found a significant positive relationship between all
types of involvement Again this finding replicates Cho and Boster’s (2005) results. This
means that the stronger one is involved with their political attitudes (value or outcome), the
more they care about how their opinionsare being perceived. Although this relationship was
significant, this study failed to find any patterns of information processing clearly associated
with impression-relevant involvement. One explanation for this non-finding is that a factor
may be confounding this relationship (which is demonstrated by the lack of conelation with
integrative complexity). This factor is potentially whether impression-relevant involvement is
interacting with involvement strength or attitude extremity. This study did not measure
whether, or how, integrative complexity and involvement type relates to attitude extremity.
Therefore, people who hold strong, or extreme, political attitudes may be dealing with more
public scrutiny or conformity issues than people not holding as outwardly strong or extreme
attitudes. Thus, the true nature between the relationship of information processing and
impression-relevant involvement remains to be determined. This relationship merits
investigation because on a college campus political nonrrs are quickly transmitted and
potentially influential. Understanding with more clarity how norms effect message
perception could be an important tool for future campaign managers seeking the college
demographic.

The third and fourth hypotheses dealt with how integrative complexity motivates
patterns of information processing. Specifically, the predictions involved whether people
higher in integrative complexity would view the initial two messages as being less

contradictory, and whether these participants would shift their opinions less after viewing the

51

contextual information. The reason why integrative complexity should have negatively
related to message contradiction was that people who are high in this construct have a
firndamental need to integrate, rather than differentiate, inforrrration. This need should have
been especially salient in this study because it was mentioned that the same politician made
the statements. There were two different ways contradiction was measured. One scale asked
the participant three items all gauging the degree to which the two messages were in
contradiction with one another. The second was a single item measure that asked the
participants whether they thought the two items were related. It was found that there was no
relationship between integrative complexity and perceptions of contradiction; however, a
relationship did exist between integrative complexity and the second measme. Participants
higher in integrative complexity stated that the messages were more closely related to one
another than participants low in integrative complexity. This was a logical relationship
because people higher in integrative complexity are more able to integrate different
information than people low in integrative complexity. One implication of this finding is that
people higher in integrative complexity have more information to draw upon when creating
their political attitudes because they are more able to find relationships among differentiated
pieces of information, whereas peOple low in integrative complexity may tend to rely on
singular pieces of information when forming their attitudes.

It is important to mention that degree of initial contradiction and integrative
complexity were heavily influenced by whether the source of the message was a preferred
source or an Opposing source. This initially seemed a bit paradoxical with the thinking that
people higher in integrative complexity should be less susceptible to these biases than people

low in integrative complexity. The fact that this was not the case however, perhaps re-

52

emphasizes the strong relationship between outcome-relevant and value-relevant
involvement. In this hypothesis it was found that when a democrat viewed a Democratic
message, those higher in integrative complexity perceived the messages as being far less
contradictory than Republicans high in integrative complexity who found the messages to be
somewhat contradictory. This pattern was similar in the case of Republicans receiving
republican messages. Here too directional bias was evident in that the more integratively
complex participants viewed the messages as less contradictory than members high in
complexity but from an opposing party. Furthermore, there were differences in these
interpretive patterns based on the issue. This pattern of information processing was more
prevalent for the affirmative action issue than the capital punishment issue. This could be
because either the messages were not equivalently contradictory from issue to issue, or that
the issues triggered different information processing techniques in the individuals. One
reason for this is that affirmative-action was a recently hot issue heavily debated at the
university because it was recently voted on, and overturned, by the state government. This
cunent event could have made the context and findings more idiosyncratic to this public
university than the issue of capital punishment.

The two predictions made in hypothesis five concerned how outcome-relevant
involvement would relate to perceptions of initial contradiction, and opinion change
following the contextual information. It was thought that people high in outcome-relevant
involvement would demonstrate patterns similar to people who scored high in integrative
complexity. This is because people motivated by accmate information processing goals
would likely bring in outside information to reconcile any immediate contradictions. Thus,

people high in involvement would scrutinize the message pair more thoroughly, and

53

consequently would perceive less contradiction than those not motivated by similar outcome
related goals. Findings from this study revealed that when both issues were analyzed in
aggregate, this relationship was not found to be the case. When splitting up the data by issue,
however, a negative association occurred in the case of capital punishment between outcome-
relevant involvement and perception of initial contradiction. This suggests that participants
who were more outcome-involved with capital punishment perceived less contradiction. For
the issue of affirmative action, this association was not observed. Although not statistically
significant, there was a slight positive correlation between outcome-relevant involvement and
initial message contradiction. This could mean that for this issue, more involvement led to a
higher perception of contradiction. Again, the reason why this discrepancy by issue occurred
is an empirical question that merits further investigation.

Another factor in the relationship between outcome-relevant involvement and initial
message contradiction was political party. The capital punishment issue showed that
Democrats who were high in outcome-relevant involvement found the two messages as being
far less contradictory than participants reporting no political affiliations. This again suggests
that although outcome-relevant involvement is conceptualized as being more accuracy
driven, patterns of directional thinking are still evident in the case of message perception and
the need to reconcile discrepant information.

The second part of this hypothesis posited that outcome-relevant involvement should
also be negatively related to opinion change after the presentation of contextual information.
The reason behind this assertion was that people high in this involvement would have already
scrutinized the information, and any new information brought forth would therefore not be as

closely depended upon when processing the messages. It should be mentioned that this

54

hypothesis did not state that change would not be possible, but rather acknowledges that
small pieces of information would not be as heavily relied upon for participants stronger in
this involvement type. Participants not knowing as much about politics, or the issue, would
likely be more susceptible to information change and information source than participants
who are more involved. Despite this reasoning however, findings failed to support this
hypothesis. It was found that opinion change, and perceptions of contradiction following the
contextual information were relatively unaffected by outcome-relevant involvement. One
post-hoe explanation for this is that much of the reasoning through the contradiction was
done before the presentation of new information. Because this reasoning was reflected in the
variance of scrutiny potentially given to the initial connadiction, new information did not
really affect what anyone later thought about the messages. An alternative explanation is that
perhaps this pattern demonstrates participants need to report political opinion consistency.
Regardless of these speculations, an area that definitely merits further research is how
contextual information is perceived. The dependent measures present in the current study
were not exhaustive enough to yield any patterns that could lead to a comfortable
understanding regarding the effectiveness of the reconciling information.

In hypothesis five it was found that the relationship between outcome-relevant
involvement and message perception was affected by party affiliations. Although this was
not foreseen because it was thought that this type of involvement would be relatively
I unaffected by preferential bias, in the case of value-relevant involvement party affiliation was
considered. All parts of hypothesis six asserted that patterns of message perception,
specifically initial contradiction and opinion change, would relate to value-relevant

involvement in different ways depending on whether the source was from a favored or

55

opposing political party. It was thought that when a politician was from a favored political
party, participants higher in value-relevant involvement would view the messages as less
contradictory than participants lower in this type. The explanation for this is basically that in
an attempt to preserve one’s ego, it becomes important to view messages from favored
politicians in a positive way. Because contradiction has gotten bad publicity in recent
political times, it makes sense here that participants would view their political party of choice
as being less contradictory than their Opposition. In the case Of the favored politician, it was
found that consistent with the above logic, participants did rate their own candidates as being
less contradictory than an opposing source. The pattern predicted for opposing candidates,
however, was not statistically significant. This finding merits further investigation because it
suggests that while ego-defense does firnction to promote and support one’s favored opinion,
it does not simultaneously outwardly disparage or discredit the Opposition’s opinion in this
scenario. It would have made sense for participants to see opposing messages as more
contradictory, however this was not the case in a significant way. In fact, the data show that
at times people are more accurate or objective when viewing an opposing party message than
a favored party message. This was not a predicted finding, but an empirical question that
natrnally follows is why this would be the case.

Hypothesis six also argued that degree of opinion change based on degree of value-
relevant involvement would also be influenced by political affiliation. For the same ego-
defensive reasons, it was thought that when the source was favored the participant would be
more likely to buy into the contextual information, and would shift their Opinion even more
in support Of their affiliation. Although this finding was not quite significant, it was found to

be the case that people higher in value-relevant involvement did not significantly change their

56

Opinion upon receiving new information. This suggests that when information is revealed
about one’s already preferred political party the participant will likely not change their
opinion very much because they essentially have already been convinced. While the
predicted relationship was consistent for the favored source, findings from participants who
viewed an Opposing candidate told a different story. It was initially suggested that value-
relevant involvement would be positively related to degree Of Opinion change for an opposing
source. It was found however that no relationship existed between these two variables. This
suggests that for an opposing source, contextual information may not change any Opinions
and participants may not pay much attention to new information when it concerns a party in
which they are not affiliated.

Many of the previous hypotheses have analyzed involvement type separately, and
then correlated the dependent variables with the strength Of a particular type of involvement.
One of the arguments advanced at the beginning Of this paper, however, was that people who
were involved would perceive messages differently than people low in involvement.
Therefore, it was important to look at general involvement and to label participants as high or
low tO see if there are differences in message perception. When doing this, it was found that
participants low in involvement perceived messages as being more contradictory than
participants high in involvement. This is explainable because, as previously mentioned,
affiliation breeds bias so when people are unaffiliated through non-involvement, their
judgments appear to be either more objective or more heuristic. This relationship was
especially strong in the case Of reported Democrats for capital punishment. It was found that
Democrats low in involvement perceived the messages as more contradictory than

Democrats high in involvement.

57

Hypothesis 7 also asserted that people who are highly involved would shift their
opinion less than people who are not very involved. It was speculated that this would be the
case because people low in involvement would be more likely to accept information as is,
without much scrutiny or outside information brought in to evaluate the veracity Of the
message. This would be indicative of mOre heuristic processing. Further, it was thought that
people low in involvement would see the contextual information as more accurate than
people high in involvement. Findings from this study revealed, however, that neither Of these
relationships were consistent with results. There were no significant differences based on
involvement for Opinion change after the contextual information or accuracy judgments
regarding this information. This suggests that level of general involvement does not
necessarily explain how people make sense of different pieces Of political information. One
idea is that the reason no change occurred and no accuracy judgments were made is because
people low in involvement are generally apathetic about politics. Thus, the credibility Of the
information made no difference. This claim however, is only speculative and again more
attention should begin to be paid concerning how people interpret potentially reconciling
information.

Finally, while different relationships were predicted and reported for type of
involvement and perceptions of the political messages, nothing was predicted for how
impression-relevant involvement would affect perception. Impression-relevant involvement
was not predicted to be tOO salient or influential in this study because no one was present to
assess the participant’s political Opinions. Regardless, the final research question examined
whether people who scored high in impression-relevant involvement would show significant

differences in their perception Of messages from a favored versus an opposing source. It was

58

found that high levels Of impression-relevant involvement did not effect the participants’
reporting of initial contradiction or Opinion change based on the source of the message. The
only difference that approached significant was that participants high in impression-relevant
involvement viewed opposing party’ 5 contextual information as more accurate than that from
their favored party. Because this finding is a bit unusual, the reason why this would be the
case remains to be determined.

Aside fi'om some methodological implications already discussed in this section,
results from this study also allow for inferences to be made regarding the political attitudes of
college students and the sophistication Of current political messages. Based on findings
regarding integrative complexity, involvement, and message perception, it was demonstrated
that political sophistication effects who is paying attention to messages, and how closely an
individual may scrutinize or make sense Of the content within the message. This suggests
that current campaign tactics that attack the Opposing candidate by exposing past political
contradictions may be a relatively ineffective strategy. The reason implied by current
findings is that people who are already affiliated with a political party will likely bias
information about their own candidate rendering the message ineffective, and will pay less
attention to messages concerning Opposing candidates because exposing this politician’s
weakness needs no affirmation. The participants in this study who were most integratively
complex were the participants claiming no political affiliations. These participants also
viewed the messages as being generally less contradictory regardless Of the affiliation
ascribed to the message. The lack of partisan bias suggests that participants reporting no
affiliation reason that political messages are not outright contradictory because they occur

within a context. They are able to not let their partisanship clout their judgment, and also use

59

their integrative reasoning to use outside information to make sense Of this contradiction. In
the current study, more than one third Of the participants reported no affiliation. Ifthis
finding generalizes to other campuses, this implies that there are many votes to be gained if
these Independents can be convinced by a persuasive rhetorical appeal. It must be kept in
mind however that if Independents are, on average, demonstrating higher levels Of political
reasoning, campaign messages must become more complex to appeal tO this group Of
individuals. This implies that current strategy, which targets the masses through relatively
simple campaigns, are only polarizing the already decided, while failing to engage the
Independent voter seeker noteworthy and credible information.

An implication that follows from this suggestion is that college students, despite their
reputation, appear politically capable of making informed decisions. Although many
students in this study provided a disclaimer stating, “I don’t know much about politics,”
almost every participant was capable of forming coherent and accurate political Opinions that
reflect knowledge of the political system and affiliations in this country. The reason students,
despite evidence to the contrary, may believe that they are politically unaware is because
campaigns neglect their Opinion, which sends the message that students do not understand
politics in an important way. Because politics appears intentionally confirsing, many students
with valid viewpoints may innocently dismiss themselves from the political process. If more
Of an effort was made at harnessing the Opinion students already possess, perhaps political
participation for this demographic would not be so disappointing.

Another global implication from this study is the complexity Of using contradicting
messages as a campaign strategy. Although initial predictions concerning how integrative

complexity, involvement, and political affiliations, affected perceptions of initial

60

contradiction were relatively straightforward, what happened after participants’ exposme to
contextual information became a lot more ambiguous. Political research in campaigning
often uses Opinion or attitude change as the main dependent variable; this study suggests
however that there is a lot going on in the perception Of the initial messages that must be
better understood before making claims about how to create a persuasive message. If
audiences are better targeted and more thoroughly researched campaigns may ultimately be
more effective in understanding how the functions served by initial attitudes and values allow
for message salience, retention, and scrutiny. The argument advanced in this paper is that
using college students as a directly targeted, and economically plausible demographic may be
the margin Of error needed tO win an election.

Although this study began to look at how political attitudes effect message
perceptions, there are limitations that may have affected results and merit replication. The
first limitation is the small sample size analyzed for many of the hypotheses. When initially
setting the number of participants, it was difficult tO predict how many would report each
political affiliation. Further, it was difiicult to match issue, affiliation, and message condition
in ways that increased each cell size without more resources and participants to draw upon. It
was also unexpected how many students would report no affiliation. Although this finding
was not surprising post-hoe, it was not anticipated in the operationalization stage. Another
unexpected finding hindering the study’s sample size was the large differences present by
issue. Initially it was thought that affinnative-action and capital punishment would have
yielded analogous results. Thinking this would be the case would have allowed for bigger

cell sizes in each message condition. The fact that this was not the case however, has

61

interesting implications for the greater study of political communication, because in this
study issue type did moderate many results.

Another inherent limitation in the study was the measurement Of integrative
complexity. Although precautions were made to get the coders reliable to the developers of
the integrative complexity scoring manual, and then to one another, the difficulty Of the
construct may have led to measurement issues that would not be detectable.

One measurement that could have been added to this study would have been
participant’s Opinion on the issue itself. Although this measurement was intentionally left out
in an effort to stay true to integrative complexities assertion that structure Of thought should
not be confused with content Of Opinion, knowing this information may have been
explanatory in the interpretation of some unexpected findings. It was previously mentioned
how Opinion extremity may have been confounded with measurements Of some forms Of
involvement, and while this should not have affected measures Of integrative complexity, it
might have affected message perception in ways that the current design could not measure.

Finally, there were not many findings that were significant when using reports of the
contextual information. It was thought that the items should scale into three dimensions
being: accuracy, consistency, and bias; however, items gauging this were not scaling together
as well as they should have. This suggests that there was more going on with perceptions Of
the contextual information that the questionnaire could not measure. It is suggested that
future research include more items so that the presence Of this contextual information could
be better understood and could perhaps help explain how people use new information to

make sense of their already existing political judgments.

62

Future research in the area of political communication, individual attitudes, and
perception Of political messages should answer some of the empirical questions interspersed
throughout this section. More could be understood with regard to who is paying attention to
what types Of political messages, who is capable Of Opinion change and at what level of
message complexity can this change occur, how methodological choices effect end results,
and lastly how norms function to induce political attitudes in a place like a college campus.

This study began to explore the development of college students’ attitudes regarding
political issues and how they perceive political messages. It is important that research
continue in this vein to understand how individuals make sense Of the political world around
them. Especially in such a highly populated area with so many similar people such as a
college campus, political research can serve to understand how public Opinion can flourish in
such environments. Currently, most registered voters are adults who have voted in several
elections. The voters who consistently participate are the people who are more difficult to
persuade. College students however are embarking upon the political process for the first
time. Although some relegate to their parent’s affiliation when making these political
choices, some look for answers fi'orn their peers and from the media on tough political issues.
Ifresearchers can begin to explore the depth and breadth of these Opinions, more could be

understood about what messages can help win elections.

63

APPENDICES

64

APPENDIX A
INTEGRATIVE COMPLEXITY CODING SCHEME
(Baker-Brown et al., 1992; Golec, 2002; Rosenberg, 1988)

Sequential Level (scores 1 & 2)
Interviewers were classified as possessing thought reflective Of this level if they revealed an
ability to:
0 Observe, notice, remember, and recall concrete Objects, events, people, etc.;
o Organize Observed and/or remembered events in a sequence based on loose and
subjective associations;
0 Opinion usually expressed as absolute or concrete and usually highly evaluative;
End result is Often the imposition of a dichotomous category structtne;
Usually qualification without elaboration, understands other viewpoints exist but
does not consider much more about the explanation behind other viewpoints;
o Demonstrates conditional acceptance Of other perspectives.

Advanced Sequential Level (Score of 3)
Interviewers were classified as possessing thought reflective Of this level if they revealed all
the above-mentioned Operations, and others allowing for more abstract reasoning such as:
0 Form simple relationships between concrete Objects, events within a reported
sequence;
Fractionate, analyze concrete Objects, events, etc.;
Recognizes alternative perspectives and accepts these as being relevant or
legitimate;
0 Can articulate multiple alternatives and perspectives and begins to understand
their relationship;
0 Probability statements are evident and ambiguity is more tolerated.

Linear Level (Scores 4 & 5)
Interviewers were classified as possessing thought reflective Of this level if they revealed all
the above-mentioned Operations, and others allowing for more abstract reasoning such as:
o Generalize relations between concrete Objects, events, etc.;
o Relate Objects and events to each other, abstract rules, laws, and properties from
relations between concrete Objects, events, etc.;
0 Evaluate Objects, events, etc., based on authority and social norms, conventions,
and standards understood as stable, absolute, and externally given;
Judgment is Often withheld and tension between alternatives is acknowledged;
Integration begins to be expressed through superordinate statements that reveal
interactivity and interdependence between alternate perspectives;
o Causal attributions are made that explains differences between perspectives.

Advanced Linear Level (Score of 6)

65

Interviewers were classified as possessing thought reflective Of this level if they revealed all
the above-mentioned Operations, and others allowing for more abstract and complex
reasoning such as:

Construct relationships between abstracted rules, laws, and properties;
Decentrate — see more than one-at-a-time aspect of Obj ects, events, etc., consider
simultaneously more than one possible point Of view on a problem;

Evaluate Objects, events, etc., based on recognition Of relativity of social norms,
based on individual choice according to preferences, without way Of determining
what is better or worse; recognition Of difference between moral (should be) and
practical (is);

Comparison Of outcomes is considered between alternative courses Of action;

Systematic analysis and hypothesis testing becomes possible as means for
information searches.

Systematic Level (Score of 7)

Interviewers were classified as possessing thought reflective Of this level if they revealed all
the above-mentioned operations, and others allowing for more abstract and complex
reasoning such as:

Generate hypotheses about novel objects, evens, and relations by juxtaposition Of
the abstract rules, laws, and properties;

Reflect on process Of generating judgments about Objects, events, and relations;
Construct general model Of system Of abstract rules, laws, properties, and
relations; use it as a frame of reference in analyzing and interpreting Objects,
events, relations, etc.;

Coordinate points Of view, define relations between different aspects of Objects,
events, etc., and between different perspectives on a problem, seeing some
aspects in context of others;

Evaluate Objects, events, etc., based on recognition of relativity Of social norms,
according tO given flame of reference that provides ways of choosing between
better and worse; awareness Of relativity and context of this choice; recognition Of
difference between moral (should be) and practical (is); attempts to integrate.

An overarching principle or perspective pertaining to the nature Of the
relationship between alternatives;

Hierarchical integration is evident and complex trade-Offs are made and
understood between conflicting goals.

66

APPENDD( B
INVOLVEMENT SCALES

Impression-relevant involvement

1. Talking about my political beliefs has little effect on what others think Of me. (Reverse
coded) ,

2. The impressions that others have Of me are very much affected when I talk with them
about my position on political issues.

3. The kind Of Opinion that I express in public about politics has little efiect on what others
think of me. (Reverse coded)

4. People may judge me on the basis of the Opinion that I express in public about political
issues.

5. IfI express the right kind Of Opinion about political issues, people will find me more
attractive.

Outcome-relevant involvement

My thoughts regarding political issues has little impact on my life. (Reverse coded)

All in all, the effect of politics on my life is small. (Reverse coded)

My life changes are a direct result of politics.

Laws created because of political issues have little effect on me. (Reverse coded)

It is easy for me to think of ways that political issues affect my life.

It is difiicult for me to think Of ways that political issues impact my life. (Reverse coded)
All in all, the effect of political changes in my life would be little. (Reverse Coded)

My Opinions on political issues are relevant to my daily life.

Political issues have direct consequences to my life.

PPNQMPPNT‘

Value-relevant involvement

1. The values that are most important to me are what determine my stance on political issues.
2. Knowing my positions on political issues is central tO understanding the kind Of person I
am.

3. My political position has little to do with my beliefs about how life should be lived
(Reverse coded)

4. My political position is based on the values with which I try to conduct my life.

5. The arguments for or against political issues are relevant to the core principles that guide
my life.

6. My beliefs about how I should live my life determine my political position.

7. My political positions reflect who I am.

8. My political Opinions do not reflect who I am. (Reverse coded)

9. My political positions are of much personal importance tO me.

67

APPENDIX C
QUESTIONAIRRE ITEMS

Demographics Questionnaire
1. How Old are you?
2. Are you male or female?
3. Which state are you from?
4. What is your ethnicity/race?
Political Attitudes
Voting Behavior
1. Have you ever voted before?
2. If so, how many times and for which election:
- Local elections _
- National elections
3. DO you intend to vote in the next local election?
4. DO you intend to vote in the next national election?
Political Ideology
3. What is your political affiliation?
a. Democrat
b. Republican
c. Independent
(1. Green Party
e. Other
f. I don’t consider myself to have any political affiliations
4. What would you say is your strength of political party identification?
Weak12 3 4 5 67Strong
5. What would you say is your strength Of political Opinions?
Weak12 3 4 5 6 7Strong
6. I would consider myself to be conservative
Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree
7. I would consider myself tO be liberal
Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree
8. What is your mother’s political affiliation?
a. Democrat
b. Republican
c. Independent
(1. Green Party
e. Other ‘
f. None
g. Don’t know
9. What is your father’s political affiliation?
a. Democrat f. None
b. Republican g. Don’t know
c. Independent e. Other
d. Green Party

68

APPENDIX D
MESSAGE EVALUATION QUESTIONNAIRE
Dependent Measures and Manipulation Checks

Perceived Contradiction (Time 1 and Time 2)

1. Too what degree do you feel that the two previous messages contradicted each other?
Not at All Contradicting 1 2 3 4 5 6 7 Strongly Contradicting

2. The two statements presented disagree with one another.

Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree

3. The two statements are not at all contradictory.

Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree

4. The two statements have nothing to do with one another.

Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree

Perceptions Of new information

1. The new information presented is accurate.

Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree

2. The two initial messages are in contradiction with one another.
Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree

3. I think the new information accmately explains how the first two messages relate.
Strongly Disagree 1 2 3 4 S 6 7 Strongly Agree

4. The new information has nothing to do with the initial two messages.
Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree

5. The new information is inaccurate.

Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree

6. The new information is biased

Strongly Disagree 1 2 3 4 5 6 7 Strongly Agree

Manipulation Checks
1. The source Of this message is a
a. Democrat
b. Republican
0. Independent
(1. Not sure

69

APPENDIX E
QUESTIONNAIRES BY ISSUE

Impression-relevant involvement (Issue A: Affirmative-Action)

1. Talking about my beliefs concerning affirmative-action has little effect on what others
think of me. (Reverse coded)

2. The impressions that others have of me are very much affected when I talk with them
about my position on affirmative action.

3. The kind of Opinion that I express in public about affirmative-action has little effect on
what others think Of me. (Reverse coded)

4. People may judge me on the basis of the opinion that I express in public about
affirmative action.

5. If I express the right kind of opinion about affirmative-action, people will find me
more attractive.

Outcome-relevant involvement (Issue A: Affirmative-Action)

1. My thoughts regarding affirmative—action has little impact on my life. (Reverse coded)
2. All in all, the effect of affirmative-action on my life is small. (Reverse coded)

3. My life changes are a direct result Of affirmative-action legislation.

4. Laws created because Of affirmative-action have little effect on me. (Reverse coded)
5. It is easy for me to think of ways that affirmative-action legislation affect my life.

6. It is difficult for me to think Of ways that affirmative-action impact my life. (Reverse
coded)

7. All in all, the effect Of legal affirmative-action changes in my life would be little.
(Reverse Coded)

8. My Opinions on affirmative-action are relevant to my daily life.

9. Affirmative-action rulings have direct consequences to my life.

Value-relevant involvement (Issue A: Affirmative-Action)

1. The values that are most important to me are what determine my stance on
affirmative-action.

2. Knowing my positions on affirmative-action is central tO understanding the kind of
person I am.

3. My position on affirmative-action has little tO do with my beliefs about how life
should be lived. (Reverse coded)

4. My position on affirmative-action is based on the values with which I try to conduct
my life.

5. The arguments for or against affirmative-action are relevant to the core principles that
guide my life.

6. My beliefs about how I should live my life determine my position on affirmative-
action.

7. My position on affirmative-action reflects who I am.

8. My position on affirmative-action does not reflect who I am. (Reverse coded)

9. My stance on affirmative-action is Of much personal importance to me.

70

Impression-relevant involvement (Issue B: Capital Punishment)

1. Talking about my beliefs concerning the death penalty has little effect on what others
think Of me. (Reverse coded)

2. The impressions that others have Of me are very much affected when I talk with them
about my position on the existence Of the death penalty.

3. The kind Of opinion that I express in public about death-penalty has little effect on
what others think Of me. (Reverse coded)

4. People may judge me on the basis of the Opinion that I express in public about capital
punishment.

5. If I express the right kind of opinion about capital punishment, people will find me
more attractive.

Outcome-relevant involvement (Issue B: Capital Punishment)
My thoughts regarding the death penalty has little impact on my life. (Reverse coded)
All in all, the effect of the death penalty on my life is small. (Reverse coded)
My life changes are a direct result of capital punishment legislation.
Laws created because of capital punislunent have little effect on me. (Reverse coded)
It is easy for me to think Of ways that capital punishment legislation affects my life.
. It is difficult for me to think of ways that capital punishment impacts my life.
everse coded)
7. All in all, the effect of changes to the death penalty laws in my life would be little.
(Reverse Coded)
8. My Opinions on capital punishment are relevant to my daily life.
9. Capital punishment rulings have direct consequences to my life.

awewwr

Value-relevant involvement (Issue B: Capital-Punishment)

1. The values that are most important to me are what determine my stance on capital
punishment.

2. Knowing my positions on the death penalty is central to understanding the kind Of
person I am.

3. My position on capital punishment has little to do with my beliefs about how life
should be lived. (Reverse coded)

4. My position on capital punishment is based on the values with which I try to conduct
my life.

5. The arguments for or against capital punishment are relevant to the core principles that
guide my life.

6. My beliefs about how I should live my life determine my position on the death
penalty.

7. My position on capital punishment reflects who I am.

8. My position on capital punishment does not reflect who I am. (Reverse coded)

9. My stance on capital punishment is of much personal importance to me.

71

APPENDIX F
MESSAGES USED IN THE STUDY

The message posted first in all cases was the message typically associated with the source
of the messages party affiliation

Issue A: Affirmative-Action

Pro: " We need to move away from a world of lingering biases and towards one in which
Opportunity is equal. Affirmative action has been good for America, when done right, it is
flexible, it is fair, and it works. The point isn’t to give out jobs, but instead to provide
opportunities to prove one self. Whether you are a female or a minority candidate, doors
close. Affirmative action can Open these doors of opportunity.”

Anti: “I feel that people of all races must be treated equally under the law, and we know
that society has not fully achieved that idea]. Yet, as we work to address the wrong Of
racial prejudice, we must not use means that create another wrong, and thus perpetuate
our economic divisions. Quota systems that use race to include or exclude people from
higher education and the Opportunities it Offers are divisive, unfair and increasingly
impossible tO justify.”

Exculpatory Statement: This politician made this second statement 4 years after the
initial statement, in a lower-middle class White community that historically struggles
with getting students tO higher education. The second statement addressed a growing
concern that state Universities were considering ethnicity before socioeconomic status,
which perpetuates class inequality. This speech was made to motivate non-minority,
low-income families, to pursue higher education.

Issue B: Capital Punishment

Pro: “Along with almost three-fourths Of the American public, I believe in capital
punishment. I believe that there are some defendants who have earned the ultimate
punishment our society has to Offer by committing murder with aggravating
circumstances present. In my view, society has not only the right, but the duty to act in
self defense to protect the innocent.”

Anti: “The way some states administer the death penalty constitutes "cruel and unusual
punishment.” The death penalty is used disproportionately against the poor, who cannot
afford expensive legal counsel, as well as racial, ethnic and religious minorities. Further,
as in this case, the death penalty was applied arbitrarily and inconsistently, which is
evidenced because wrongly convicted, innocent people have received death penalty
sentence, and tragically, were killed by the state.”

Exculpatory Statement: This politician made this second statement in a low-income
neighborhood near a state penitentiary, 4 years after the initial statement. The second

72

statement addressed recurring accusations that states were not abiding by capital
punishment in the way the law intended, because research had recently indicated that 50
innocent people were wrongly executed at this penitentiary in particular.

73

APPENDIX G
Table 1
Correlations Between Integrative Complexity and Outcome—Relevant Involvement by

Political Party and Issue Type

 

 

Issue All Participants Democrats Republicans Other NO
Affiliation

Both .198* .211 -.161 .191 .432"
Issues (N = 122) (n = 39) (n = 25) (n = 12) (n = 46)
Affirmative- .260* .290 -.1 14 .192 .636"
Action (n=6l) (n=22) (n=15) (n=7) (n= 17)
Capital- .130 .398 -.267 -.350 .l 11
Punishment (n = 61) (n = 17) (n = 10) (n = 5) (n = 29)

 

*p < .05. ”p < .01.

74

Table 2

Correlations Between Integrative Complexity and Value—Relevant Involvement by

Political Party and Issue Type

 

Issue All Participants Democrats Republicans

 

Other No

Affiliation
Both . 193 * .266 .069 .199 .229
Issues (N = 122) (n = 39) (n = 25) (n = 12) (n = 46)
Affirmative- .349** .426* .152 .391 .599“
Action (n = 61) (n = 22) (n = 15) (n = 7) (n = 17)
Capital- -.060 .169 -.187 -.075 -.185
Punishment (n = 61) (n = 17) (n = 10) (n = 5) (n = 29)

 

*p < .05. "p < .01.

75

Table 3
Correlations Between Outcome-Relevant Involvement and Perceived Message

Discrepancy by Political Party and Issue Type

 

 

Issue All Participants Democrats Republicans Other NO
Affiliation
Both -.062 -.262 .062 -.140 .000
Issues (N = 122) (n = 39) (n = 25) (n = 12) (n = 46)
Affirrnative- .146 . -.023 .338 .226 .1 17
Action (n = 61) (n = 22) (n = 15) (n = 7) (n = 17)
Capital- -.287* -.744** -.027 -.627 -.084
Punishment (n = 61) (n = 17) (n = 10) (n = 5) (n = 29)

 

*p < .05. "p < .01.

76

Table 4
Correlations Between Value-Relevant Involvement and Perceived Message Discrepancy

by Political Party, Message Condition, and Issue Type

 

Issue Democrats Republicans

 

Democrat Message Condition (n = 31)

Affirmative- -.157 .527
Action (n = 10) (n = 7)
Capital- -.766* -.802
Punishment (n = 8) (n =6)

 

Republican Message Condition (n = 32)

Affirmative- .146 -.735*
Action (n = 12) (n = 8)
Capital- -.287* .451
Punishment (n = 8) (n = 4)

 

*p < .05. I""‘p < .01.

77

REFERENCES

Baker-Brown, G., Ballard, E. J ., Bluck, S., de Vries, B., Suedfeld, P., & Tetlock, P. E.
(1992). The conceptual/integrative complexity scoring manual. In C. P. Smith, J.
W. Atkinson, D.C. McClelland, & J. Veroff (Eds), Motivation and Personalig:
imdbookgf them_ati_c content a_n_zdvsis (pp. 401-418). New York: Cambridge
University Press.

Battistich, V. A., & Aronoff, J. (1985). Perceiver, target, and situational influences on
social cognition: An interactional analysis. Journal of Personality and Social
Psychology, 49, 788-798.

Burdein, I., Lodge, M., & Taber, C. (2006). Experiments on the automacticity Of
political beliefs and attitudes. Political Psychology, 27, 359-371.

Caprara, G. V., Barbaranelli, C., & Zimbardo, P. G. (1999). Personality profiles and
political parties. Political Psychology, 20, 175-197.

Chaiken, S., Liberman, A., & Eagly, A. H. (1989). Heuristic and systematic processing
within and beyond the persuasion context. In J. S. Uleman & J. A. Bargh (Eds),
Unintended thought: Limits of awareness, attention, and control (pp. 212-252).
New York: Guilford.

Cho, H. & Boster, F. J. (2005). Development and validation Of value-, outcome, and
impression-relevant involvement scales. Communication Research, 32, 235-264.

Cho, I. (2005). Media, interpersonal discussion, and electoral choice. Communication
Research, 32, 295-322.

Cohen, G. L. (2003). Party of policy: The dominating impact of group influence on
political beliefs. Journal of Personality and Social Psychology, 85, 808-822.

Conover, P. J. (1988). The role Of social groups in political thinking. British Journal of
Political Science, 18, 51-76.

Corning, A. F ., & Meyers, D. J. (2002). Individual action toward engagement in social
action. Political Psychology, 23, 703-729.

Crosby, F. J. (1976). A model Of egoistical relative deprivation. Psychological Review,
83, 85-113.

Dolan, K. A., & Holbrook, T. M. (2001). Knowing versus caring: The role Of affect and
cognition in political perceptions. Political Psychology, 22, 27-44.

78

 

Duncan, L. E. (2005). Personal political salience as a self-schema: Consequences for
political information processing. Political Psychology, 26, 965-976.

Eliasoph, N. (1998). Avoiding politics: How Americans produce apathy in everyday life.
New York: Cambridge University Press.

Festinger, L. (1957). A Theory Of Cognitive Dissonance. Evanston, IL: Row, Peterson.

Foster, M. D., & Matheson, K. (1995). Double relative deprivation: Combining the
personal and political. Personality and Social Psychology Bulletin, 21, 1167-
1177.

George Mason University. (n.d.). United states elections project. Retrieved October 14,
2006, from Dr. Michael McDonald’s Web site: http://elections.gmu.edu/

Golec, A. (2002). Cognitive skills as predictor Of attitudes toward political conflict: A
study of polish politicians. Political Psychology, 23, 731-757.

Greenberg, J., & Jonas, E. (2003). Psychology motives and political orientation- The left,
the right, and the rigid: Comment on J ost et al. (2003). Psychological Bulletin,
129, 376-382.

Greenberg, A. (2003). New generation, new politics. The American Prospect, 14.
Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.

Huddy, L. (2001). From social to political identity: A critical examination of social
identity theory. Political Psychology, 22, 127-156.

Jenkins, J. C. (1983). Resource mobilization theory and the study of social movements.
Annual Review Sociology, 9, 527-5 53.

Johnson, B. T., & Eagly, A. H. (1989). Effects of involvement on persuasion: A meta-
analysis. Psychological Bulletin, 106, 290-314.

Johnson, B. T., & Eagly, A. H. (1990). Involvement and persuasion: Types, traditions,
and evidence. Psychological Bulletin, 107, 375-384.

lost, I. T., & Maj or, B. (2001). The psychology of legitimacy. New York: Cambridge
University Press.

Katz, D. (1960). The functional approach to the study Of attitudes. The Public Opinion
Quarterly, 24, 163-204.

Lee, T. (2002). Mobilizing public opinion: Black insurgency and racial attitudes in the
civil rights era. Chicago, IL: University Of Chicago Press.

79

Lodge, M., & Taber, C. S. (2005). The automacticity of affect for political leaders,
groups, and issues: An experimental test of the hot cognition hypothesis. Political
Psychology, 26, 455-482.

Petty, R. E., & Cacioppo, J. T. (1986a). Communication and persuasion: Central and
peripheral routes to attitude change. New York: Springer-Verlag.

Petty, R. E., & Cacioppo, J. T. (1986b). The elaboration likelihood model of persuasion.
In L. Berkowitz (Ed.), Advances in experimental social psychology (V 01. 19, pp.
123-205). San Diego, CA: Academic Press.

Putnam, R. (1996). The strange disappearance of civic America. The American
Prospect, 7.

Rokeach, M. (1960). The open and closed mind New York, NY: Basic Books, Inc.

Rosenberg, S.W. (1988). Reason, ideology, and politics. Princeton, NJ: Princeton
University Press.

Rudolph, T. J. (2006). Triangulating political responsibility: Motivated formation of
responsibility judgments. Political Psychology, 27, 99-122.

Schaller, M., Boyd, C., Yohannes, J ., & O’Brien, M. (1995). The prejudiced personality
revisited: Personal need for structure and formation of erroneous group
stereotypes. Journal of Personality and Social Psychology, 68, 544-555.

Schroeder, H. M., Driver, M. J ., & Streufert, S. (1967). Human information processing.
New York: Holt, Rinehart, and Winston.

Sidanius, J. (1978). Cognitive functioning and socio-political ideology: An explorative
study. Perceptual and Motor Skills, 46, 515-530.

Sidanius, J. (1988). Political sophistication and political deviance: A structural equation
examination of context theory. Journal of Personality and Social Psychology, 55,
37-51.

Skitka, L. J ., & Tetlock, P. E. (1993). Providing public assistance: Cognitive and
motivational processes underlying liberal and conservative policy preferences.
Journal of Personality and Social Psychology, 65, 1205-1223.

Streufert, S., & Streufert, S. (1978). Behavior in the complex environment. Washington,
DC: Winston.

Tajfel, H. (1981). Human groups and social categories. Cambridge: Cambridge
University Press.

80

Tetlock, P. E. (1983). Cognitive style and political ideology. Journal of Personality and
Social Psychology, 45, 118-126.

Tetlock, P. E. (1986). A value pluralism model of ideological reasoning. Journal of
Personality and Social Psychology, 5 0, 819-827.

Tetlock, P. E., Peterson, R. S., & Lerner, J. S. (1996). Revising the value pluralism
model: Incorporating social context and context postulates. In C. Seligman, J.
Olson, and Zanna, M. (eds.), The Ontario Symposium: Vol. 8. The Psychology of
Values (pp. 25-51). Mahway, NJ: Erlbaum.

Thomsen, C. J ., Borgida, E., & Lavine, H. (1995). The causes and consequences of
personal involvement. In R. E. Petty & J. A. Krosnick (Eds), Attitude strength:
Antecedents and consequences (pp. 191-214). Mahwah, NJ: LEA.

Turner, J. C., Hogg, M. A., Oakes, P. J ., Reicher, S. D., & Wetherell, M. S. (1987).
Rediscovering the social group: A self-categorization theory. Oxford: Blackwell.

United States Census Bureau. (2006, March). Voting and registration in the election of
November 2004. Retrieved October 14, 2006, from http://www.census.gov/prod/
2006pubs/p20-556.pdf

Van Hiel, A., & Mervielde, I. (2003). The measurement of cognitive complexity and its
relationship with political extremism. Political Psychology, 24, 781-801.

Webster, D. M. (1993). Motivated augmentation and reduction of the overattribution
bias. Journal of Personality and Social Psychology, 65, 261 -271 .

Westen, D., Kilts, C., Blagov, P., Harenski, K., & Hamann, S. (in press). The neural
basis of motivated reasoning: An fMRI study of emotional constraints on
political judgment during the US. Presidential election of 2004. Journal of
Cognitive Neuroscience.

Zimbardo, P. G. (1960). Involvement and communication discrepancy as determinants
of opinion conformity. Journal of A bnormal and Social Psychology, 60, 86-94.

81