Two of your friends have asked you to describe the professor that you had for American history last semester. To one friend, you describe the professor using two extremely positive traits: brilliant and witty. To the other friend, you describe the professor using two extremely positive traits, but you also use one trait that is only moderately positive: prompt, brilliant, and witty. Which of your friends will form the most favorable attitude toward the professor?
The theme of this chapter is that a person's attitude about some person, object, or issue is determined by the information the person has about the stimulus and by how that information is combined or integrated to form one overall impression. In our example, the information consists of the traits that have been associated with the professor. But how is this information put together? If the information is added, then the friend who heard the three positive traits would have a more favorable attitude than the friend who heard the two traits, but if the information is averaged, then the friend who heard only the two extremely positive traits would have the most favorable attitude. In this chapter we will focus first on the different types of information that we can have about people, objects, and issues; and then we will present three different approaches to the study of attitude change that emphasize how people combine information. (We will return to our adding vs. averaging dilemma later in the chapter in box 7.3).
Fishbein and Ajzen ( 1975) have distinguished among three different types of information that we may have about attitude objects. The first kind of information associates an attribute with an object on the basis of direct personal experience and has been called a primitive belief ( Rokeach, 1968) or a descriptive belief ( Fishbein & Ajzen, 1975). Since these beliefs comprise our basic truths about the world, attitudes based on descriptive beliefs are very difficult to change. Also, as we saw in chapter 1, attitudes based on descriptive beliefs (i.e., attitudes based on direct contact with the object of judgment) correlate highly with behaviors toward that object ( Fazio & Zanna, 1981). A second kind of information consists of beliefs that go beyond directly observable events and are called inferential beliefs. For example, if you believe that Alice runs faster than Barbara on the basis of direct observation, and that Barbara runs faster than Clara on the basis of direct observation, you would likely infer that Alice runs faster than