Human beings have a natural tendency to react with some degree of positive or negative affect to any object or concept of psychological significance (Fishbein & Ajzen 1975; Eagly & Chaiken 1993). We like or dislike certain people, support or oppose various policies, regard some activities as pleasant and others as unpleasant, have favorable views of certain television programs but unfavorable views of others, and so forth. The expectancy value (EV) model is concerned with the origins and structure of these “social attitudes” (Feather 1982).
According to the EV model, the overall evaluation of or attitude toward an object is a function of the information or “beliefs” we have about the object in question. In the domain of communication research, the EV model has been applied to predict viewer attitudes and viewer exposure to various types of television programs and shows: news in general (Palmgreen & Rayburn 1982; Babrow & Swanson 1988) and health related news reports in particular (Cooper et al. 2001), soap operas (Babrow 1989), and prime-time entertainment programming (Galloway & Meek 1981). In recent years, it also has been applied to predict computer technology use of Canadian elementary and secondary teachers from schools in Quebec (Wozney et al. 2006).
Assumptions
The EV model assumes that in the course of our daily lives our experiences lead us to acquire different beliefs about various objects, actions, and events. These beliefs may be formed as a result of direct experience with the object (e.g., watching a new Steven Spielberg film); they may be acquired indirectly by accepting information from other people or from printed and electronic media about the new Spielberg film; or they may be self-generated through inference processes based on knowledge stored in memory about Steven Spielberg films in general.
Each belief associates the attitude object with an attribute. For example, we may come to believe that a newspaper (the object) is created by a team of competent journalists and covers new and relevant information that is presented in an unbiased way. Because the attributes that come to be associated with the object are already valued positively or negatively, we simultaneously and automatically acquire an attitude toward the object. Specifically, the subjective value of each attribute contributes to the attitude in direct proportion to the strength of the belief, i.e., to the subjective probability that the object has the attribute in question. The way in which beliefs about an object (o) combine to produce an overall attitude is shown in the following equation:
Formula 1
It can be seen that the strength of each belief (bio), i.e., the subjective probability that object o is characterized by attribute i, is multiplied by the evaluation of the associated attribute (ei), and the resulting products are summed over the n beliefs about the object. According to the EV model, a person’s attitude toward an object (Ao) is a function of this summative belief index. In this fashion, we learn to like objects we associate with largely desirable characteristics, to hold unfavorable attitudes toward objects we associate with mainly undesirable characteristics, and to form ambivalent attitudes toward objects we associate with both desirable and undesirable characteristics.
Of course, individuals are not expected actually to perform the mental calculations specified in the equation. The EV model is taken not as an accurate description of the way in which attitudes are formed but rather it is assumed that the attitude formation process can be modeled as if individuals were performing the stipulated calculations. It is also important to realize that people can form many beliefs about any psychological object, but that only a relatively small number is readily available in memory at any given moment. It is these readily accessible beliefs that are assumed to be the prevailing determinants of a person’s attitude.
Identifying Accessible Beliefs
Different approaches exist to identify accessible beliefs about an attitude object. Two popular approaches are the thought listing technique and the use of focus groups. For the thought listing technique, individuals are asked to list in a free response format any positive and any negative aspects of the object that come readily to mind. In focus groups, potential users of, say, a mass medium (e.g., television program, newspaper) are brought together in small groups and, in a permissive atmosphere under the guidance of a moderator, discuss various aspects of the media offering in question.
Based on one of these techniques in most applications of the EV model, the most frequently listed responses (thought listing technique) or most frequently mentioned beliefs from group discussions (protocol from focus groups) are selected to construct a list of modal accessible beliefs, i.e., beliefs that are common in the population of interest. Once a list of accessible attributes has been constructed, a new sample of participants is asked to rate the likelihood and the valence associated with each attribute. That is, they are asked to rate how likely it is that the object has the attribute (belief strength) and to rate the attribute on an evaluative scale (attribute evaluation). These two ratings are multiplied, and the products are summed in accordance with the above equation. In addition, as a direct measure of attitude, participants are also asked to rate the attitude object itself on an evaluative scale. In accordance with the EV model, empirical research has demonstrated strong correlations between this direct attitude measure and the summed belief × evaluation index.
In short, in the EV model it is assumed that our beliefs form the informational foundation for our attitudes. Although often quite accurate, beliefs can be biased by a variety of cognitive and motivational processes. They may be irrational, based on invalid or selective information, be self-serving, or otherwise fail to correspond to reality. However, no matter how they were formed or how accurate they are, beliefs represent the information we have about the world in which we live, and they form the cognitive foundation for our attitudes toward aspects of that world.
Application Of The Model
The EV model can be used not only to account for the formation and structure of attitudes but also to help explain behavioral decisions. People form attitudes not only toward physical objects, institutions, social groups, events, and mass media offerings, but also toward behaviors. Thus, we may hold favorable or unfavorable attitudes toward watching a television program, reading a newspaper, playing a computer game, using a certain computer technology for teaching purposes, and so forth.
When the object of the attitude is a behavior, the relevant beliefs that determine the attitude are readily accessible beliefs about the consequences of the behavior. For these behavioral beliefs, again, a list of modal accessible behavioral outcomes must be constructed, and participants are asked to rate the likelihood that the behavior will produce each outcome and to rate the valence of each outcome on an evaluative scale. Belief strength and outcome evaluation ratings are multiplied and the products are summed to produce the EV composite, which is again found to correlate well with a direct measure of attitude toward the behavior. Some research has focused on the question of whether the EV composite can be decomposed by factor-analytic methods into belief sub-groups representing different functions that a certain behavior serves. In a study on watching soap operas among college students, Babrow (1989) identified four positively correlated EV subscales: perceived opportunities for social interaction, anticipated learning, anticipated general amusement and entertainment, and expected opportunity for romantic fantasy.
The EV model also has a number of implications for attitude change (Ajzen & Fishbein 1980; Fishbein & Ajzen 1981). Change in an attitude can be produced by changing the beliefs that are already accessible for the recipients or by making new beliefs accessible. According to the EV model, to change an attitude in a favorable direction, the summed EV products of the beliefs that underlie the attitude must become more positive than prior to the influence attempt, whereas a change in an unfavorable direction is produced by a negative change in the summed products.
To illustrate, if a person seeks information about basketball and believes that a certain television sports channel would supply a great deal of relevant information, the individual may seek such information from that sports channel. If the individual receives basketball information at a lower level than expected, it should reduce belief strength with regard to this attribute and as a consequence change the attitude in a negative direction. If, however, the same sports channel offers information about football at a higher level than expected (and this information is also positively valued by the individual), the resulting increase in belief strength can compensate for the decrease in the former belief strength, leading in sum to an increase in the product sum and to a more positive attitude toward watching that sports channel. According to the EV model, changes in beliefs can occur either via changes in the belief strengths of the attributes or via changes in the evaluations of the attributes, or in both terms. Empirically, persuasive attempts trying to change belief strengths tend to be more effective than attempts trying to change belief evaluations, probably because evaluations of attributes are well anchored in prior learning.
Methodological Issues
A methodological issue of the EV model concerns the scaling of belief strengths and outcome evaluations that are multiplied in EV products. Usually, respondents make their ratings on 7-point scales for both belief strengths (e.g., likely vs unlikely) and outcome evaluations (e.g., good vs bad). Derived from the conventional assumption that evaluations have a bipolar quality ranging from negative to positive with a neutral midpoint, bipolar scaling of outcome evaluations is used (from –3 to +3). Controversial, however, is the most adequate scaling for the belief strength ratings. Either bipolar scaling (from –3, very improbable to, +3, very probable) or unipolar scaling (from 0, very improbable, to +6, very probable) has been used. The bipolar scaling method allows respondents to express the falsity of modal accessible beliefs (Eagly & Chaiken 1993) and enables strong disbelief in very negative attributes to make a large positive contribution to the product-sum (–3 × –3).
To illustrate, perceiving it as very unlikely (–3) that a television program will show pornographic material (–3) can contribute positively to the composite score. However, unipolar scaling may appear more reasonable given the interpretation of the belief strength as a subjective probability. Unipolar scaling requires disbelief in negative attributes to make either no contribution (0 expectancy: 0 × –3) or a small negative contribution (1 expectancy: 1 × –3) to the composite. Most studies explaining mediarelated attitudes by the EV model have employed the combination of bipolar scaling for outcome evaluations and unipolar scaling for belief strength ratings.
References:
- Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice Hall.
- Babrow, A. S. (1989). An expectancy-value analysis of the student soap opera audience. Communication Research, 16(2), 155 –178.
- Babrow, A. S., & Swanson, D. L. (1988). Disentangling antecedents of audience exposure levels: Extending expectancy-value analyses of gratifications sought from television news. Communication Monographs, 55(1), 1–21.
- Cooper, C. P., Burgoon, M., & Roter, D. L. (2001). An expectancy-value analysis of viewer interest in television prevention news stories. Health Communication, 13(3), 227–240.
- Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Fort Worth, TX: Harcourt Brace.
- Feather, N. T. (ed.) (1982). Expectations and actions: Expectancy-value models in psychology. Hillsdale, NJ: Lawrence Erlbaum.
- Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Reading, MA: Addison Wesley.
- Fishbein, M., & Ajzen, I. (1981). Acceptance, yielding and impact: Cognitive processes in persuasion. In R. E. Petty, T. M. Ostrom, & T. C. Brock (eds.), Cognitive responses in persuasion. Hillsdale, NJ: Lawrence Erlbaum, pp. 339 –359.
- Galloway, J. J., & Meek, F. L. (1981). Audience uses and gratifications: An expectancy model. Communication Research, 8, 435 – 450.
- Palmgreen, P., & Rayburn, J. D. (1982). Gratifications sought and media exposure: An expectancy value model. Communication Research, 9, 561–580.
- Wozney, L., Venkatesh, V., & Abrami, P. (2006). Implementing computer technologies: Teachers’ perceptions and practices. Journal of Technology and Teacher Education, 14(1), 173 –207.