34.4 Attitudes and Behavior

In 1969, Wicker reported on a review of 42 experimental studies that assessed attitudes and then included an observation of related behaviors. Wicker found few studies where the correlation between attitudes and behavior were as high as .30 (r = .30). The average correlation was about .15 (r = .15). Wicker concluded that "taken as a whole, these studies suggest that it is considerably more likely that attitudes will be unrelated or only slightly related to overt behaviors than that attitudes will be closely related to actions" (p. 65).

The impact of Wicker's review was immediate. By the early 1970s most social psychologists had readily accepted the negative verdict about the attitude-behavior link. Most felt that attitudes had little importance and direct relation to actions. Also during this time frame, many studies were conducted and reported that examined the impact of behaviors on attitudes, rather than the other way around. Festinger's (1957) dissonance theory was very popular, and its emphasis on the influence of behaviors on attitudes seemed to make it difficult for researchers to believe that the opposite link could be strong or even stronger.

Simonson's (1977) study discussed earlier showed that attitudes could be changed predictably by following dissonance theory guidelines, but that the impact of attitudes on achievement was not easily identified. Simonson's results demonstrated a one standard deviation improvement in attitude scores, but no significant improvement in related achievement scores. This study was one of many reported during this period which supported the position taken by many that attitude and behavior had little relationship to each other. Most research of the 1970s dealing with attitudes did not attempt to demonstrate a direct link between attitudes and behavior (Eagly & Chaiken, 1993).

Since then, however, there has been a reexamination of the attitude-behavior link, and a resurgence of interest in this area has occurred. As a matter- of fact, the generalization that attitudes do not predict behavior is now considered inaccurate (Eagly & Chaiken, 1993). Parenthetically, as one reads the debate about attitudes and behaviors that raged in the social psychology literature during the 1970s and 1980s, it is interesting to compare it to the ongoing debate in instructional technology between those who say "media will never influence achievement' (Clark, 1983, 1994) and those who take a more moderate approach to the relationship between mediated instruction and achievement. The similarities are interesting. In the attitude-behavior debate, subsequent research has shown that neither extreme position was correct, and the following section of this chapter will present what currently is known about the attitude-behavior link.

This section will include a brief explanation of the approaches taken recently by attitude researchers and will concentrate on theories of social psychology. The section will conclude with a discussion of attitudes and instructional behaviors.

First, while Wicker's (1969) premise about the weak link between attitudes and behavior gained widespread acceptance, many took issue with his study's methods. Wicker reviewed only a narrow sample of studies that were heavily weighted toward laboratory research, Many outstanding studies were not examined by Wicker, and several writers who reviewed survey research reported that this literature showed a moderately strong relationship between attitudes and behaviors (Kelman, 1974; Schuman & Johnson, 1976). This critique of Wicker led to a series of new proposals about, and new examinations of, the relationship between attitudes and behavior.

First, researchers began to examine attitudes as being related to an aggregate of behaviors. It was found that relatively high attitude-behavior correlations were obtained by comparing a general attitude (e.g., attitude toward chemistry or attitude toward homework) to a measure with an aggregate of attitude-relevant behaviors (e.g., taking chemistry courses, talking about chemistry in the study hall with friends, using chemistry examples in other classes).

Fishbein and Ajzen (1975) reported that attitudes typically predicted multiple-act criteria better than single-act criteria. They generalized that attitudes and behaviors must be compatible to ensure a strong relationship. In other words, general attitudes are good predictors of general behaviors (e.g., attitudes toward affirmative action are predictors of actions related to affirmative action), and specific attitudes, especially attitudes toward behaviors, are good predictors of specific actions (attitudes toward studying chemistry predict nicely the act of studying). Unfortunately, many researchers had examined, and still are examining, quite general attitudes (e.g., attitudes toward chemistry), and quite specific behaviors (e.g., achievement on a chemistry test).

Fishbein and Ajzen maintained that a consistent terminology was needed. In other words, specific attitude constructs should be identified if specific actions are to be correlated to them. An attitude toward studying chemistry every evening could be measured by an attitude test, and subsequent studying of chemistry could be determined by asking students' parents to keep a journal of their child's study habits. Fishbein and Ajzen predicted that the correlation between this kind of specific attitude and specific behavior would be quite high.

Fishbein's efforts in this area produced a model that he and Ajzen called the theory of reasoned action. It is now considered an excellent model of the psychological processes that explain observed links between attitudes and behaviors. The theory of reasoned action suggests that the cause of behavior is a person's intention to engage in the behavior. Attitudes influence behavior by their influence on intentions, which are decisions to act in a particular way. The issue of how an attitude was transformed into action was resolved by adding another psychological event, the formation of an intention. Intention was explained to be the person's motivation to exert effort to carry out a behavior.

This theory was popular since it had an inherent reasonableness about it. People were assumed to behave as they intended to behave. They were theorized to act in ways that allowed them to obtain favorable outcomes and to meet the expectations of others. The theory of reasoned action can be summarized as follows:

a. Behavior is determined by the intention to engage in the behavior.

b. Intention is determined by attitude toward the behavior and the subjective norm to which the attitude is related.

c. Attitude is determined by behavioral beliefs and evaluation of the likely outcomes of a behavior.

d. Subjective norms are determined by the normative beliefs of the person and the motivation to comply with the relevant actions.

Many believe that this theory provides a complete theory of voluntary behavior. Critics have indicated that they do not consider the theory of reasoned action to be a general theory of behavior. Rather, it is considered by them to be a theory of the immediate causes of voluntary action.

In part because of criticisms, Ajzen proposed an alternative theory of planned behavior that attempted to enlarge the Fishbein-Ajzen model (Ajzen, 1991). Ajzen stated that for nonhabituaI behaviors that are easily executed by almost everyone without special circumstances, the theory of reasoned action was adequate. When behaviors are more difficult to execute, and when a person needs to take control over needed resources in order to act, the theory of planned behavior is a better predictor of behavior than the theory of reasoned action. In the theory of planned behavior, control is taken into account as a variable labeled "perceived behavioral control," which is defined as a person's perception of how easy or difficult it would be to perform the action.

Perceived control affects behavior in two ways: First~ it influences the intention to perform the behavior. Second, it may have a direct impact on the behavior itself. Ajzen proposed that people tend to engage in behaviors to the extent that they believe that they have control over the behaviors, in other words, to the extent that they have confidence in their ability to perform the behavior.

In a series of studies that examined the prediction of behavior, it was found that when perceived behavioral. control was taken into account, along with attitude toward a behavior, the average R for predicting intentions was .71 (Ajzen, 1991). Research suggests that the addition of perceived control to the model of reasoned action results in a more comprehensive model that applies to behaviors that require skills, resources, and other inputs that are not available merely because people decide to act.

From the low point of the late 19609, when many social scientists believed that attitudes were not closely related to behaviors, new information has been made available to the point where most now believe there is considerable interaction between attitudes, behaviors, and other variables. High correlations between attitudes and overt behaviors can be produced by aggregating several behaviors to create a measure that corresponds to the attitude measure. The theory of reasoned action and the theory of planned action provide direction to the study of the prediction of behavior, especially where attitudes are concerned.

Eagly and Chaiken (1993) also proposed a composite attitude-behavior model that is especially attractive because of its comprehensiveness (Fig. 34-1). This model demonstrates that behavior is likely to be partially determined by attitudes, but that the relation between attitudes and behavior is best understood by placing attitudes in the context of other factors that also help to determine behavior, such as habits, intentions, and perceived utilitarian outcomes.

Ultimately we must return to the purposes of this chapter: the design of persuasive messages that are delivered by media, and the relationship between media and attitudes. It is apparent from the literature of social psychology that a direct relationship between attitude formation and the production of educational behaviors such as achievement is not straightforward. Rather, the development of attitude positions that are desirable and planned is only one step in the process of promoting educational relevant actions. Attitudes contribute to learning outcomes, but are only one of several important variables. Arguments listed previously in this chapter provide support for the need to understand the use of mediated messages designed to persuade. Later in this chapter, a series of guidelines for producing attitude changes will be proposed.

It would be inappropriate to assume that the development of new attitude positions will directly and predictably influence educational behaviors. Rather, attitudes are one component of a system that predicts behaviors. For those interested in predicting behavior from attitudes, the literature provides guidelines. First, single general attitudes are not likely to predict general actions. At the very least, very specific attitudes and very specific behaviors should be identified for correlation. Second, general attitudes are probably related to a collection, an aggregate, of behaviors. Finally, other variables such as motivation, intention, and personality traits are intervening forces that should be considered in the attitude-behavior formula. Interestingly, even the critics of attitude-behavior research are consistent in their opinion that it is possible, even easy, to modify attitudes predictably, and that attitudes play some role in determining actions.