A comparison of three social-psychological models of attitude and behavioral plan: prediction of contraceptive behavior

We compared the predictive validities of three prominent models of attitudes and behavioral decisions: Rosenberg's instrumentality-value model, Fishbein's belief-evaluation model, and Beach's adaptation of subjective expected utility theory. Seventy female undergraduates rated each of the models' components and reported their attitudes and behavioral plans toward using three different methods of contraception. With the traditional across-subjects prediction procedure, the Rosenberg model generally accounted for 5-25% less variance in subjects' attitudes and behavioral plans than the Fishbein an Beach models, which were not different. With a within-subject prediction procedure, the Rosenberg model was again the least accurate, and the Fishbein and Beach models had similar predictive accuracy. As hypothesized, within-subject predictions were more accurate than across-subjects predictions. The relatively poor performance of the Rosenberg model was attributable to the instrumentality component. In addition, we found that the Beach model could be simplified with no appreciable loss in predictive accuracy. Finally, a subject's personal normative beliefs emerged as a strong independent predictor of behavioral plan.     

Beliefs, attitudes, and intention in the context of abortion

Women's beliefs, attitudes, and intention, defined according to Fishbein and Ajzen (1975), were studied in relation to procedures for abortion: a commonly used surgical method, vacuum aspiration, and a recent medical alternative. Subjects were 53 women requesting an abortion and 53 nonpregnant women. Effects of having had an abortion prior to the study and of situation were obtained as expected in belief ratings and attitude measures. Choice of the medical alternative was related to having distinctly opposite attitudes to methods. Choice of the surgical treatment related to having slightly negative attitudes to both methods. In the actual situation more positive attitudes had been expected and were obtained. Previous experience of abortion and situation were assumed to be external factors with effects on intention if mediated by the attitudes. A logistic regression model showed a good fit and predicted the intentional choice of treatment from attitude scores.     

Salience of Anti-Abortion Beliefs and Commitment to an Attitudinal Position: On the Strength,Structure, and Predictive Validity of Anti-Abortion Attitudes1

The role of beliefs in attitude formation and the impact of commitment to an attitude on its predictive validity were studied in the context of anti-abortion attitudes. Undergraduates (N= 152), identified as pro-choice or pro-life, expressed their beliefs, attitudes, intentions, and restriction preferences with respect to making abortion illegal. In addition, they indicated their commitment to their positions and their willingness to distribute a pro- or anti-abortion petition. Salient beliefs—identified in a pilot study—were found to predict attitudes, intentions, restriction preferences, and petition choice significantly better than nonsalient beliefs. Salient beliefs also discriminated significantly between pro-choice and pro-life respondents, providing useful information about the cognitive underpinnings of anti-abortion attitudes. In addition, accuracy of predictions increased significantly with commitment, even when attitude extremity was statistically controlled. These findings support the summation theory of attitude (Fishbein, 1963) and demonstrate the importance of attitude strength in determining the structure and predictive validity of attitudes.     

Psychological Responses to the Threat of Nuclear War

We investigated the nature of psychological responses to the threat of nuclear war and the relationships between attitudes and behavior in three studies. The first was a quasi-experimental study of the effects of a film depiction of nuclear war on attitudes and behavior. Results suggested that exposure was mediated by a sense of control over political events, but that exposure itself had no significant effect on psychological responses to nuclear war. The second was an experimental investigation of the effects of efficacy enhancing, fear arousing, or informational tactics as well as individual differences on memory for relevant information and political activism on the issue of nuclear arms control. Results indicated that men exposed to the fear arousal tactic were less likely to take action than those exposed to information alone or efficacy enhancement, whereas women exposed to fear arousal were more likely to act. Gender differences are explained in terms of the socialization of male responses to fear, the "macho" response. Those who showed less denial were more likely to act, as were those who stated intentions to take action. The third study employed structural modeling to examine the relationships between attitudes and behavior. Results supported the Model of Reasoned Action (Fishbein & Ajzen, 1975) in showing that intentions to act mediated the relationship between attitudes and behavior on this issue.     

From automatic antigay prejudice to behavior: the moderating role of conscious beliefs about gender and behavioral control

Two experiments tested whether the relation between automatic prejudice and discriminatory behavior is moderated by 2 conscious processes: conscious egalitarian beliefs and behavioral control. The authors predicted that, when both conscious processes are deactivated, automatic prejudice would elicit discriminatory behavior. When either of the 2 processes is activated, behavioral bias would be eliminated. The authors assessed participants' automatic attitudes toward gay men, conscious beliefs about gender, behavioral control, and interactions with gay confederates. In Experiment 1, men's beliefs about gender were heterogeneous, whereas women's beliefs were mostly egalitarian; men's responses supported the predictions, but women's responses did not. In Experiment 2, the authors recruited a sample with greater diversity in gender-related beliefs. Results showed that, for both sexes, automatic prejudice produced biased behavior in the absence of conscious egalitarian beliefs and behavioral control. The presence of either conscious process eliminated behavioral bias.     

Understanding the attitudes and intentions of future professionals toward self-help

Examined the attitudes, beliefs, and intentions toward self-help groups of 168 graduate students in clinical psychology and social work from five universities using the theory of reasoned action as a model (Fishbein, 1979). Participants held positive attitudes and beliefs regarding self-help and intended to collaborate. Participants who were members of self-help groups had significantly greater intentions to collaborate and had more positive beliefs vs. nonmembers. There were no differences between social work and psychology students. Path analysis showed that students who held positive attitudes and beliefs and perceived that their faculty were positive regarding self-help had intentions to collaborate with self-help groups. Involving self-help groups as partners in professional training was considered empowering and a wise use of the expert resources that groups can provide.     

Predicting behavior on a computer from intentions, attitudes, and experience

Research on computer attitudes has generally focused on describing the cognitive elements. The present study used the Ajzen and Fishbein model to predict intentions to work hard in a computer course and actual time on the computer. In addition, the influence of experience on predictions was examined. Results from 74 social science students showed that the Ajzen and Fishbein model is effective at predicting both intentions and behavior. Unlike previous findings, experience did not interact with attitudes. Finally, a short cognitive attitude scale was an additional significant predictor beyond intentions in explaining behavior variance.     

Fear or Moral Indignation? Predicting Attitudes Toward Parolees1

Instrumental beliefs, value-expressive beliefs, and attitudes related to parolees were investigated in 2 contexts. In Study I, students (N= 180) responded to value-expressive measures in a mass survey and then, a month later, read 2 scenarios and completed measures of instrumental beliefs, attitudes, and intentions. Instrumental beliefs were better predictors of behavioral attitudes than were value-expressive beliefs. In Study 2, the format of the measures was altered and all measures were administered concurrently. As in Study I, results showed that instrumental beliefs were consistent and strong predictors of attitudes toward parolees. Attitudes related to parolees appear to be based more on practical concerns (instrumental beliefs) than on moral or symbolic issues (value-expressive beliefs).     

Correspondence and mediational properties of the Fishbein model: An application to adolescent alcohol use

The generalizability of the Fishbein model for behavior prediction was extended to a new field behavior, alcohol drinking by adolescents. The model's two components (i.e., attitudes and normative beliefs) varied in their predictive importance depending on the behavior's situational context. Following this field demonstration, two issues related to properties affecting the model's predictive effectiveness were investigated. First, it was found that the predictability of alcohol-drinking intentions (BI) declined when attitudes (A_act) and normative beliefs (NB) did not correspond with BI in terms of alcohol type (beer, liquor, or wine) and drinking situation (home, party, or pub). A second experiment tested the hypothesis that exogenous variables (those other than A_act and NB) would have a direct influence on BI. A comprehensive test was provided within the context of Jessor's multivariate predictive system for adolescent alcohol use. Present data indicate only a marginal increase in explained variance occurred when applying this system, thereby generally indicating the sufficiency of the Fishbein model's two components.     

Situational factors affecting the weighting of predictor components in the Fishbein model

The Fishbein regression model (B ≈ BI ≈ (Aact)w₀ + (NBs)w₁) is designed to predict behavior (B) from attitudes (Aact) and normative beliefs (NBs). Using a variation of theAjzen and Fishbein (Journal of Experimental Social Psychology, 1970,6, 466–487) method, potential limitations of the model were investigated. It was found that certain situational factors were capable of altering the nature of the model significantly. Aact carried a nonsignificant regression weight when the participant had had no prior direct experience with the act and carried a significant weight when the participant had had experience with the act. NB carried a nonsignificant weight under competitive motivational set conditions and a significant weight under cooperative set conditions. These findings raise the question of the generality of the Fishbein model, since different prediction models were required by different situational conditions. The hypothesized association between attitude and behavior appears to be dependent on prior experience with the behavior, while the association between norms and behavior seems to be dependent on consistency between personal motivation and perceived social expectations.     

Predicting Undergraduate Condom Use with the Fishbein and Ajzen and the Triandis Attitude-Behavior Models: Implications for Public Health Interventions1

The present study used two social-psychological theories of attitude-behavior relationships to investigate condom use by college undergraduates. One purpose of the present research is to provide information for public health practitioners to better design theory-based interventions to reduce the incidence of STD's by increasing the use of condoms. Another purpose is to compare the relative usefulness of two attitude-behavior models in predicting condom use. The Fishbein and Ajzen theory of reasoned action (Ajzen, 1988; Fishbein & Ajzen. 1980) was compared with a version of Triandis' theory of attitude-behavior relations (Triandis, 1977). Of major interest was whether variables in this version of the Triandis model increased the explanatory power of the Fishbein and Ajzen model in the prediction of condom use intention and behavior. In addition, specific outcome beliefs and normative beliefs held by subjects were used to predict condom use. One hundred and ninety college undergraduates were given a questionnaire designed to measure the components of the two models. Three months after completing the questionnaire, subjects were contacted by phone to record the subject's frequency of condom use (if the subject reported having had intercourse) during the 3-month period between initial assessment and follow-up. In the prediction of intention to use a condom, results indicate that one variable from the Triandis model, personal normative beliefs, increases the explanatory power offered by the expectancy value and normative belief components of the Fishbein and Ajzen model. In the prediction of reported condom use during the 3-month follow-up period, two variables from the Triandis model—perceived susceptibility and AIDS fear—significantly increased the predictive power of the Fishbein and Ajzen model. Based on these results, the authors propose an alternative model to explain undergraduate condom use. The implications of these findings for public health interventions to encourage the use of condoms is discussed.     

Behavioral Goals for a Diverse Organization: The Effects of Attitudes, Social Norms, and Racial Identity for Asian Americans and Whites

This research identified 5 behaviors that may enhance the effectiveness of a diverse organization and used Fishbein and Ajzen's (1975) theory of reasoned action to identify predictors of these behaviors. Results from samples of White and Asian undergraduate students from 2 universities generally supported the Fishbein and Ajzen model. Attitudes and subjective norms were significant predictors of behavioral intentions when gender, race, and social desirability bias were controlled. Racial identity also had a significant, positive effect on attitudes toward diversity-related behaviors among the Asian American students, but no significant effect among Whites. These results supported our reasoning that members of historically excluded racial groups with strong racial identities will be most likely to welcome organizational attempts to become more pluralistic because pluralism means that their valued identities will be respected rather than repressed.     

Interactive effects of writing and reading a persuasive essay on attitude change and selective exposure

Subjects wrote an essay either supporting their own position on an issue or irrelevant to that issue. Subsequently, they received a communication that either supported or opposed their position, and that either claimed to present the only reasonable position or acknowledged the viability of alternative positions. Then, they reported their own beliefs in the position being advocated, and finally were given an opportunity to obtain subsequent information that either supported or opposed this position. Attitude change in the direction of a proattitudinal communication occurred only when both (a) this communication asserted that only the position advocated was defensible and (b) subjects had previously written a proattitudinal essay themselves. Moreover, either of these factors was sufficient to prevent attitude change in the direction of a counterattitudinal communication. Data supported predictions based upon a joint consideration of commitment and reactance effects. Subjects' subsequent preference for supporting over opposing information was greater when they had written a supporting essay than when they had not, greater when they had read an opposing communication than when they had read a supporting one, and greater when the communication they read was presented as stating the only reasonable position, no matter which position it advocated.     

The Instrumental and Symbolic Functions of Attitudes toward Persons with AIDS1

Five studies explored the psychological bases of attitudes toward persons afflicted with Acquired Immune Deficiency Syndrome (AIDS). These studies examined both instrumental and symbolic bases of these attitudes. In Studies 1, 2, and 3, both instrumental factors (e.g., beliefs about the probability of one's own child contracting AIDS) and symbolic factors (general attitudes toward homosexuality) independently contributed to the prediction of attitudes toward having one's child attend classes with a nonhomosexual person having AIDS. In Study 4, only attitudes toward homosexuality (symbolic factors) and not beliefs about contagiousness related to students' expressing a desire to transfer from a class with an AIDS-infected professor. In Study 5, subjects role played the situation experienced by subjects in Study 4. A wider array of instrumental concerns was assessed. While both instrumental and symbolic factors were related to attitudes of role-playing subjects, the specific instrumental concerns of importance were related to beliefs about subjects' feeling comfortable with the professor and not the contagiousness of AIDS. Thus, these results paralleled those of Study 4. These findings are discussed with regard to their relevance for understanding the varying functions of attitudes and for understanding the stigmatization of disease victims.     

The Predictive Benefits of Importance: Do Issue Importance Ratings Improve the Prediction of Political Attitudes?

Using data collected in Arizona during the 2000 Presidential election, this study explores whether expectancy value (EV) models predicting attitudes toward candidates and toward voting for candidates can be improved by incorporating measures of issue importance. More specifically, attitudes toward candidates were predicted from beliefs about the candidates' stands on 8 issues, and attitudes toward voting for the candidates were predicted from beliefs that voting for a candidate would lead to the implementation of the 8 issue positions. Ratings of the importance of the 8 issues were used to develop several different EV models. The results of our study indicate that importance ratings do not add much to the EV model's prediction of attitudes.     

Seeking Social Support in Old Age as Reasoned Action: Structural and Volitional Determinants in a Middle-Aged Sample of Argentinean Women1

Given that the availability of social support influences physical and psychological well-being, the provision of positive contacts is often crucial for older women. The purpose of the present study was to test the utility of both structural and volitional factors as determinants of support-seeking intentions. Two intentions to seek social support were studied in a group of 106 middle-aged women: (a) the intention to go to a doctor at least once a year after 55; and (b) the intention to visit a family member at least once a week after 55. The structural variables of interest were age, educational level, presence of partner, number of children, and number of grandchildren. The volitional factors assessed were intentions, beliefs, attitudes, and norms. There was little evidence that structural factors influenced support seeking. In contrast, as predicted from the theory of reasoned action (Ajzen & Fishbein, 1980; Fishbein & Ajzen. 1975), intentions to go to the doctor and to visit a family member could be successfully predicted from attitudes and norms.     

Effects of a persuasive communication on beliefs, attitudes, and career choice

Fishbein's Theory of Reasoned Action was used to formulate a persuasive communication in an attempt to influence unclassified American college students' beliefs, attitudes, intentions, and behaviors regarding signing up for a career as a registered nurse. A two-stage cluster sample was used to assign 90 male and female students to either an experimental or control group. After persuasive communication exposure, the experimental group showed a significantly more positive change in beliefs, attitudes, and intentions than did the control group exposed to a neutral message. Sign-up rate was also statistically significant for the experimental group. With the Fishbein model to predict sign-up behavior, no other scores were found to add to the prediction once behavioral intention was entered into the model. Change in behavioral intention explained 49% of the variation in behavior. Normative belief scores did not approach statistical significance.     

Predicting and understanding mothers' infant-feeding intentions and behavior: testing the theory of reasoned action

The present study examines the applicability of Fishbein and Ajzen's theory of reasoned action to the prediction and understanding of how primiparous and multiparous mothers intended to feed their infants and how they actually fed these infants during the 6 weeks following delivery. Measures of attitudes to behavior, subjective norms, and behavioral intentions were taken during the last trimester of pregnancy. Behavior was assessed by self-report 6 weeks postpartum. In most respects the findings supported the theory of reasoned action. However, attitudes to behavior were found to make an independent and significant contribution to the prediction of infant-feeding behavior, and the previous behavior of multiparous mothers explained an independent and significant proportion of variation in their behavioral intentions. The relative importance of the attitudinal and normative components of the theoretical model tended to vary according to whether the mothers had direct experience of the criterion behavior. Further analysis revealed that mothers who breast-fed during the 6-week postpartum period differed from those who bottle-fed exclusively during this period on a number of behavioral beliefs, outcome evaluations, and normative beliefs, and on one measure of motivation to comply. The implications of these findings for the theory of reasoned action are discussed.     

Beliefs and Attitudes in a Gubernatorial Primary: Some Limitations on the Fishbein Model1

In order to investigate hypothesized limitations on a portion of the Fishbein expectancy-value model, a three wave panel survey was conducted during the 1978 Texas Democratic gubernatorial primary. The survey elicited respondents' beliefs about the candidates, attitudes toward the candidates, and commitment to the candidates. Analyses revealed that the correlations between evaluative beliefs about candidates (Σe_ib_i) and global attitudes toward them (Ao) were affected to varying degrees by three factors extraneous to Fishbein's model. There was evidence that belief-attitude correspondence was affected by (1) respondent education and income, (2) degree of commitment to the candidates, and (3) familiarity with the candidates (salience).     

因果判断中经验与共变信息的结合及各自作用

研究使用相继呈现信息的方法控制了被试获得信息的顺序,从而将经验和共变信息在因果判断中所起到的作用分离了出来。结果表明：（1）个体综合两种信息进行因果判断的过程既不是简单的相加操作,也不是使用经验信息控制共变信息的进入,而是先判断经验信息,再判断共变信息是否与之一致,当出现不一致的情况时又重新考虑经验信息。（2）在改变先前判断的过程中,经验信息所起的作用更大,其中又以当其证明待判断原因不可信时所产生的改变更大。