THE COGNITIVE ORGANIZATION OF BELIEFS ABOUT BEHAVIORS
BY
ANNE BC DURAN
A dissertation submitted to the Graduate School
in partial fulfillment of the requirements
for the degree
Doctor of Philosophy
Major Subject: Psychology
New Mexico State University
Las Cruces, New Mexico
December 2003
1
“The Cognitive Organization of Beliefs About Behaviors,” a dissertation prepared by Anne Duran, in partial fulfillment of the requirements for the degree, Doctor of Philosophy, has been approved and accepted by the following:
Linda Lacey
Dean of the Graduate School
David Trafimow
Chair of the Examining Committee
Date
Committee in charge:
Dr. David Trafimow, Chair
Dr. W. Larry Gregory
Dr. Laura J. Madson
Dr. Cookie Stephan
Dr. Walter G. Stephan
ACKNOWLEDGEMENTS
Thank you to my parents, Jan and Roy Fausnaugh, for their continued and consistent support. Thank you for all of your emergency trips and all of your supportive words.
Thank you to Walter Stephan. You know how they say “without whom this would not have been possible”? I wish I could communicate to you how much your support has meant to me, especially during those lowest of low times. Thanks. You are my hero.
Thanks to Lausanne Renfro-Fernandez—without whom this would not have been possible!
VITA
January 10, 1961Born in Russell, Kansas
1979Graduated from Teen Mothers High School,
Arvada, Colorado
1995Bachelor Degree in Psychology,
Metropolitan State College of Denver,
Magna Cum Laude
1998Master of Arts Degree in Psychology,
New Mexico State University
2001 - PresentLecturer, California State University, Bakersfield
Professional and Honorary Societies
Society for Personality and Social Psychology
Western Psychological Association
American Psychological Society
American Psychological Association, Div. 2, Society for the Teaching of Psychology
Publications
Duran, A., & Trafimow, D. (2000). Cognitive organization of favorable and unfavorable beliefs about performing a behavior. Journal of Social Psychology, 140, 179-187.
Gregory, W.L., & Duran, A. (2000). Scenarios and Acceptance of Forecasts. In J.S. Armstrong (Ed.), Principles of Forecasting: A Handbook for Researchers and Practitioners, Kluwer Academic Press, New York, NY.
Renfro, C.L., Duran, A., & Stephan, W.G., & Clason, D. L. The role of threat in attitudes toward affirmative action and its beneficiaries. Manuscript submitted for publication.
Stephan, W.G., Diaz-Loving, R., & Duran, A. (2000). Integrated threat theory and intercultural attitudes: Mexico and the United States. Journal of Cross-Cultural Psychology, 31, 240-249.
Trafimow, D., & Duran, A. (1998). Some tests of the distinction between attitude and perceived behavioral control. British Journal of Social Psychology, 37, 1-14.
Duran, A. (1998). The Effect of Trait Type on Recall. Unpublished master's
thesis, New Mexico State University, Las Cruces.
Presentations
Allen, T., Cole, D., & Duran, A. (2002, April). Comparing measures of attitudes toward outgroup members. Poster session presented at the meeting for the Western Psychological Association, Irvine, CA.
Duran, A., Finlay, K., Stephan, W.G., & Trafimow, D. (2000, April, June). The Relationship Between Prejudice and Discrimination. Paper presented at the meeting for the Graduate Symposium, Las Cruces, NM (May). Poster session presented at the Rocky Mountain Psychological Association (RMPA), Tucson, AZ (April) and the American Psychological Society, Miami, FL (June).
Duran, A., & Madson, L. (2000, June). Teaching the Psychology of Sexual Orientation. Poster session presented at the meeting for the Rocky Mountain Psychological Association, Tucson, AZ (April) and the American Psychological Society Institute on the Teaching of Psychology, Miami, FL.
* 1st place winner, CTUP/STP Teaching Issues Competition
Duran, A., & Stephan, W.G. (1999, June). Perceptions of Threat: Predicting Attitudes Toward Affirmative Action and its Beneficiaries. Poster presented at the Graduate Symposium, Las Cruces, NM (May). Poster session presented at the American Psychological Society, Denver, CO.
* Winner of the Sigma Xi Best Poster Research Award
Duran, A., Waller, M., & Madson, L. (1999, June). Perceptions of Sexual Orientation: A Hierarchically Restrictive Trait. In K. T. Schneider and P. Rhadakrishanan (Co-Chairs), Theoretical perspectives on attitudes toward and perceptions of lesbians and gay men. Symposium conducted at the meeting of the American Psychological Society, Denver, CO (June).
Duran, A., & Trafimow, D. (1998, April). Attitude and Perceived Behavioral Control: Is There a Difference? Poster presented at the Western Psychological Association, Albuquerque, NM.
Duran, A. (1998, April). The Effect of Trait Type on Recall. Paper presented at the meeting for the Dialogue between Aggies and Miners, El Paso, TX, and the Graduate Research and Arts Symposium, Las Cruces, NM.
Duran, A., & Trafimow, D. (1997, April). Some Tests of the Distinction Between Attitude and Perceived Behavioral Control. Paper presented at the Dialogue between Aggies and Miners, Las Cruces, NM.
Duran, A., & Trafimow, D. (1996, April). Cognitive Organization of Positive and Negative Beliefs. Paper presented at the Dialogue between Aggies and Miners, El Paso, TX.
Teaching Experience
California State University, Bakersfield August 2001-present
New Mexico State University, Las Cruces Fall 1998–Spring 2000
Field of Study
Major field:Psychology
ABSTRACT
THE COGNITIVE ORGANIZATION OF BELIEFS ABOUT BEHAVIORS
BY
ANNE BC DURAN, M.A.
Doctor of Philosophy
New Mexico State University
Las Cruces, New Mexico, 2003
Dr. David Trafimow, Chair
According to the theory of reasoned action, the best predictor of whether a person will perform a behavior is the person’s intentions to perform that behavior. Researchers have found that when forming intentions to perform a behavior, people access their beliefs in clusters: beliefs that are similar to each other are grouped together. One category of beliefs is affective and cognitive; another is for and against the behavior. Affective and cognitive clustering has consistently been demonstrated, but for and against clustering is less consistent. This set of studies tests four predictions regarding the occurrence of for and against clustering. The first study examined whether the specificity of the behavior was influential in the clustering effects. The second study tested whether salience of different selves was a factor. The third study explored whether the level of morality associated with different behaviors was a factor, and the fourth study investigated the direction of the behavior. Within the four studies, consistent patterns of for and against clustering were not found; however, clustering of affective and cognitive tended to occur. Possible explanations for the inconsistent results are offered.
TABLE OF CONTENTS
Page
LIST OF TABLES...... xii
THE COGNITIVE ORGANIZATION OF BELIEFS
ABOUT BEHAVIORS...... 1
Cognitive Organization of Beliefs...... 6
The Replication Experiment...... 12
EXPERIMENT ONE: GENERAL V. SPECIFIC BEHAVIORS...... 15
Experiment One: Method...... 17
Pilot Study...... 17
Materials and Methods...... 18
Experiment One: Results...... 22
For and Against Clustering...... 22
Affective and Cognitive Clustering...... 24
Pleasantness Ratings...... 27
Experiment One: Discussion...... 27
EXPERIMENT TWO: THE SALIENCE OF THE INGROUP...... 32
Experiment Two: Materials and Method...... 36
Experiment Two: Results...... 37
For and Against Clustering...... 38
Affective and Cognitive Clustering...... 41
Proportions of Beliefs...... 43
Ratings of Beliefs...... 44
Sexual Activity...... 45
Experiment Two: Discussion...... 45
EXPERIMENT THREE: THE MORALITY ISSUE...... 50
Experiment Three: Method...... 54
Pilot Study...... 54
Materials and Methods...... 58
Experiment Three: Results...... 58
For and Against Clustering...... 59
Affective and Cognitive Clustering...... 61
Proportions of Beliefs...... 64
Experiment Three: Discussion...... 65
EXPERIMENT FOUR: FRAMING EFFECTS...... 69
Experiment Four: Materials and Method...... 71
Experiment Four: Results...... 71
For and Against Clustering...... 71
Affective and Cognitive Clustering...... 73
Sexual Activity...... 74
Experiment Four: Discussion...... 76
GENERAL DISCUSSION...... 78
Individual Differences...... 84
Implications and Applications...... 87
CONCLUSION...... 88
REFERENCES...... 89
LIST OF TABLES
Table Page
1.Pleasantness Ratings for General and Specific Behaviors.... 19
2.Clustering of For and Against Beliefs, General and
Specific Behaviors...... 23
3.Clustering of Affective and Cognitive Beliefs, General
and Specific Behaviors...... 26
4.Clustering of For and Against Beliefs After Priming...... 39
5.Clustering of Affective and Cognitive Beliefs After Priming.. 42
6.Morality Ratings and T-Tests for Matched Behaviors...... 56
7.Clustering of For and Against Beliefs, Non-Moral and
Moral Behaviors...... 60
8.Clustering of Affective and Cognitive Beliefs, Non-Moral and
Moral Behaviors...... 62
9.Clustering of For and Against Beliefs, Condom Use and
Protected Sex Behaviors...... 72
10.Clustering of Affective and Cognitive Beliefs, Condom Use and
Protected Sex Behaviors...... 75
11.Summary of For/Against Clustering and Affective/Cognitive
Clustering Across All Studies...... 81
1
THE COGNITIVE ORGANIZATION OF BELIEFS ABOUT BEHAVIORS
The theory of reasoned action is used to predict and explain behaviors (Ajzen & Fishbein, 1980; Fishbein, 1979; Fishbein & Ajzen, 1975). According to this theory, the best predictor of whether a person will perform a behavior is the person’s intentions to perform that behavior. Intentions to perform the behavior are influenced by two conceptually independent factors: subjective norms and attitudes about the behavior. Subjective norms (SN) are determined by beliefs about what important others think the person should do about the behavior, or normative beliefs (n), and the person’s motivation to comply with those beliefs (m). If the person believes that most of their important others think that person should perform the behavior, and if the person cares what those others think, this perceived social pressure will influence behavioral intentions in favor of performing the behavior:
SN nimi
Attitudes (A) are determined by the person’s beliefs about the consequences of the behavior, or behavioral beliefs (b),and the evaluation of these consequences (e). If the benefits of the behavior outweigh the costs, the attitude regarding the behavior will be generally favorable, and this positive attitude will influence behavioral intentions in favor of performing the behavior:
A biei
Researchers who examine the theory of reasoned action typically use multiple regression analyses to determine the relative contribution of subjective norms and attitudes on different behavioral intentions. For some behaviors, the beta weight for the subjective norm is higher than the beta weight for the attitudes, indicating that the relative contribution of the subjective norm component is larger than the attitude component. These behaviors are said to be under normative control (NC). For other behaviors, the attitude beta weight is larger than the subjective norm beta weight, indicating that the relative contribution of the attitude component is larger than the subjective norm component. These behaviors are said to be under attitudinal control (AC). Finlay, Trafimow, and Jones (1997) and Finlay, Trafimow, and Moroi (1999) found that many health-related behaviors tend to be NC; for these behaviors, people’s beliefs about what most of their important others think they should do have substantial influence on their intentions. Trafimow and Finlay (2001) found that attitudes had larger beta weights than subjective norms across thirty different behaviors; for these behaviors, the consequences of the behavior tended to be particularly influential on their intentions.
While some behaviors tend to be NC and some tend to be AC, the relative influence of subjective norms and attitudes on intentions can change, depending on the situation and on the person. Priming is an example of a situational influence. Ybarra and Trafimow (1998) found that primes can influence the relative beta weights of subjective norms and attitudes. Using private- and collective-self primes, they found that people exposed to the private-self prime placed more weight on attitudes than subjective norms in forming their intentions, and people exposed to the collective-self prime placed more weight on subjective norms than attitudes in forming their intentions. The level of perceived risk in a given situation can also influence whether a behavior is NC or AC: Trafimow and Fishbein (1994) and Stasson and Fishbein (1990) found that in low-risk situations, intentions to wear seat belts are AC; in high-risk situations, intentions to wear seat belts are NC. Additionally, there is evidence that individuals also may be more or less AC or NC. Trafimow and Finlay (1996, 2001) and Finlay et al. (1999) found that most individuals are under AC: the correlation between attitudes and intentions is higher than the correlation between subjective norms and intentions. However, a significant minority of participants (18% to 34%) was under NC. For NC participants, the correlation between subjective norms and intentions was higher than the correlation between attitudes and intentions.
One area of research involving the theory of reasoned action is directed toward finding other beliefs, besides normative and behavioral beliefs, which influence intentions. For example, Azjen’s theory of planned behavior (1988) expanded the theory of reasoned action by including control beliefs — beliefs regarding whether performing a behavior is under the actor’s volitional control. Control beliefs influence perceived behavioral control, which influences intentions to perform some behaviors. Other factors, such as beliefs about morals (Gorsuch & Ortberg, 1983; Manstead, 2000), affect (Triandis, 1980), confidence in normative beliefs (Trafimow, 1994), expected affect after performing the behavior (Richard, van der Pligt, & de Vries, 1996), perceived difficulty (Trafimow, Sheeran, Conner, & Finlay, 2002), affective and cognitive properties (Breckler, 1984; Breckler & Wiggins, 1989; Crites, Fabrigar, & Petty, 1994; Trafimow, Sheeran, Lombardo, Finlay, & Brown, in press) and habit (Trafimow, 2000) have also been found to influence behavioral intentions.
Another major area of research relating to the theory of reasoned action is directed toward determining how well subjective norms and attitudes predict or explain behavioral intentions. There have been several meta-analyses on the predictive ability of the theory of reasoned action or its related expansion, the theory of planned behavior (e.g., Albarracin, Johnson, Fishbein, & Muellerleile, 2001; Hausenblas, Carron, & Mack, 1997; Sheeran & Taylor, 1999; Sheppard, Hartwick, & Warshaw, 1988; Sutton, 1998). The findings in this research indicate that subjective norms, attitudes, and perceived behavioral control are good proximal predictors of behavioral intentions, and behavioral intentions are good proximal predictors of behaviors. These components explain on average between 40% and 50% of the variance in intentions, and between 19% and 38% of the variance in behavior (Sutton, 1998). The explained variance is higher when the measures of attitudes, subjective norms, and perceived behavioral control correspond to the behavior (or the behavioral intentions) regarding the target, time, context, and object (Ajzen & Fishbein, 1977; Ajzen, 1991).
Some research on the theory of reasoned action combines both the identification of different belief types and the ability of different belief types to predict intentions. In this area, researchers work to distinguish whether different belief types are distinct from each other. For example, Trafimow and Fishbein (1995) found that people distinguish between behavioral and normative beliefs, and Trafimow and Duran (1998) found that people distinguish between behavioral beliefs and control beliefs. Trafimow, Finlay, Sheeran, and Conner (2002) found that perceived behavioral control is a function of the perception of control and the perception of difficulty. They also found that control beliefs and difficulty beliefs are distinct, and each of these belief types can be independently manipulated to predict behavioral intentions. The results of these studies indicate that there are essential differences between different belief types, and different belief types may make different predictions of behavioral intentions. Understanding differences between belief types can increase the predictive ability of the theory of reasoned action.
Less attention has been directed to research regarding the manner in which different beliefs are organized and accessed to form attitudes and subjective norms. To better understand and predict behavior, it is important to begin with an understanding of how people cognitively organize and access their beliefs about behaviors. In previous research (e.g., Ybarra & Trafimow, 1998), the relative weight of attitudes and subjective norms was changed after exposure to different primes. Because attitudes and subjective norms are determined by behavioral and normative beliefs, respectively, it is possible that the primes activated different beliefs, and it was the accessing of these beliefs that influenced intentions. If the beliefs that are accessed influence intentions, it is important to investigate how these beliefs are accessed. Are there heuristics that guide which beliefs will be accessed for a particular behavior, in a particular situation? Or are beliefs randomly accessed?
According to the theory of reasoned action and related research, there are several different types of beliefs that influence intentions to behave. Investigating how these beliefs are organized can be a complex matter. The first step is to narrow the focus of attention. Generally, attitudes are the most important predictors of behavioral intentions (Ajzen & Fishbein, 1980; Finlay et al., 1999; Fishbein & Ajzen, 1975; Miniard & Cohen, 1981; Trafimow, 2000; Trafimow & Finlay, 1996; Trafimow & Finlay, 2001). The beliefs that influence attitudes, behavioral beliefs, can be classified in many different ways. For example, a behavioral belief can be either affective or cognitive (Crites et al., 1994; Trafimow & Sheeran, 1998) or for or against the behavior (Duran & Trafimow, 2000). Because attitudes tend to be the most important predictors of intentions, and because behavioral beliefs can be classified in many different ways, it would be unmanageable to examine all beliefs that influence intentions (e.g., normative, control, difficulty). The current research will be limited to two subsets of behavioral beliefs: affective/cognitive and for/against.
Cognitive Organization of Beliefs
The theory of reasoned action implies that when people make a decision about a behavior, they do so after thoroughly weighing the pros and cons of beliefs about that behavior (Thogersen, 1996). However, people have limited cognitive resources, and may not be able to consider a large set of beliefs when determining an attitude (Fishchoff, Goitein, & Shapira, 1982). It seems unlikely that all of the pros and cons, or costs and benefits, of a particular behavior are accessed each time a behavioral decision is made. Situational influences may affect the accessibility of different beliefs, and as a result, which beliefs are used to form behavioral intentions. For example, imagine a student is considering the behavior “turning on the air conditioning during studying.” The student might access beliefs such as “it will help me relax” and “it will keep me cool.” Imagine the same student considering the same behavior, but on the day that the electricity bill arrived. The belief that “electricity is expensive” may become salient, and may then be used to form attitudes and behavioral intentions.