Earlier we said that attempts at changing people’s attitudes have the ultimate goal of changing their behavior—for example, getting them to purchase a product, go on a date, or wear their seatbelts. But how well do attitudes really predict behavior?

Consider your own experience. For instance, in anticipation of hanging out with a friend who constantly mooches off you, you might think, “The next time Sven asks to borrow my car, I’m going to tell him what I really think.” Yet when Sven subsequently and predictably asks to borrow your ride, you toss him your keys with perhaps only a fleeting glance of disapproval. Some factor besides your attitude toward Sven’s mooching determined your actions. It turns out that a number of factors complicate the attitude-behavior relationship.

The first hint of such complicating factors was uncovered by the sociologist Richard LaPiere back in 1934. He traveled across the United States with a young Chinese couple, visiting 251 different hotels and restaurants. At that time, there was rather strong anti-Chinese prejudice in the United States, and there were no broadly enforced laws against ethnic, racial, or gender discrimination. LaPiere was therefore curious to see how many of the places they visited would refuse service to the Chinese couple. As it turned out, they were refused service only in one case.

What makes this finding particularly surprising is that, six months later, LaPiere wrote letters to each establishment he had visited, asking if its employees would serve Chinese individuals. Of the establishments that wrote back, 92 percent said that they would not. But of course, only one actually did refuse service. Although LaPiere acknowledged a number of problems with this study (e.g., different employees might have answered his letter than those he encountered when visiting in person), the findings were striking in their total lack of support for the notion that verbal reports of attitudes guide behavior. Decades later, the relevant literature continued to reveal surprisingly limited support for the influence of attitudes on behavior (Wicker, 1969).

One reason attitudes often are poor predictors of behavior is that sometimes people do not know what their attitudes are, or at least don’t know exactly why they feel the way they do. As we saw in chapter 5, people can easily verbalize the reasons that one might hold a particular attitude, but those reasons may not reflect their personal, gut-level feelings (Nisbett & Wilson, 1977); thus they can be poor predictors of behavior.

A second reason is that even when people do have a clear attitude that is relevant to a behavior, they often don’t appreciate that they have other relevant attitudes that pull them in other directions (Fishbein & Ajzen, 1975). For example, have you ever found yourself sitting through a movie that you initially had no interest in seeing? Although you had a negative attitude toward this movie, you may have had other attitudes, such as your desire to please your significant other, who was interested in the movie. In this example, your attitudes toward your relationship won out over your attitude toward the movie to influence your movie-going behavior. As this example foreshadows, understanding why attitudes don’t always predict behavior gives us important clues about when attitudes will predict behavior (Glasman & Albarracín, 2006).

304

Matching the Attitude to the Behavior

One reason for LaPiere’s findings is that when he asked hotel and restaurant employees whether they would serve Chinese individuals, he did not mention the specific details of his visit six months before, such as the fact that the Chinese couple was well-dressed and in the company of an American university professor. Thus, the employees’ general attitude toward serving Chinese patrons may have been too general to predict their behavior in that specific situation. This highlights an important point: Although general attitudes are poor predictors of specific behaviors, more specific attitudes fare much better; conversely, general attitudes are better predictors of more general classes of behavior.

In a classic illustration of this point, Davidson and Jaccard (1979) asked a sample of married women about their attitudes toward birth control in general. Two years later, they interviewed the women again and asked how often they had used the birth-control pill in the preceding two years (see FIGURE 8.13). In the first condition, the researchers asked women, “What is your attitude toward birth control?” They found that responses correlated .08 with Pill-taking behavior. This means that there was practically no relationship between the women’s attitude and their behavior.

Why do you think this relationship was so small? Obviously the question was quite general, glossing over many ways of thinking about the Pill. For example, people might have a positive attitude toward birth control in general but feel that the Pill is not the best method. Still others might be trying to get pregnant, so although they might advocate birth control for other people, they are not using birth control themselves.

In the second condition, the researchers asked a more specific question: “What is your attitude toward the birth control pill?” Here they found that the correlation between women’s attitude and their behavior improved to .32. This means that women who reported more positive attitudes toward the Pill were somewhat more likely to be taking it. But the relationship between attitude and behavior still was not very strong. In the third condition women were asked about their attitudes “toward using birth control pills,” and in the fourth condition, their attitude “about using the birth control pill in the next two years.” In these conditions, the correlations increased to .53 and .57, respectively. These findings show that attitudes that are more specifically relevant to a behavior are better predictors of that behavior.

305

However, in situations where you are trying to predict a general set of behaviors, more general attitudes do a much better job. For instance, if you’re trying to predict whether people will behave in environmentally friendly ways, you’re better off getting their attitudes about environmental protection in general rather than one specific attitude. For example, rather than ask them how they feel about recycling soda cans, you should measure attitudes toward a range of proenvironment practices (e.g., support for hybrid cars and protecting endangered species) and combine them into an overall environmentalism attitude (Weigel & Newman, 1976).

Self-presentational Concerns

Another reason that attitudes poorly predicted behavior in LaPiere’s study may be that when hotel and restaurant employees were responding in the moment to LaPiere and the couple’s visit, they were probably under self-presentational pressures that they were not under when they later described in a letter what they would hypothetically do. In the lobby of a hotel, for example, the employees were most likely concerned with appearing professional and avoiding an unpleasant scene, self-presentational concerns that may have influenced their behavior more than their private attitudes toward Chinese people. In the private response to a letter sent through the mail, though, they could express their negative attitudes more freely without being concerned about public scrutiny.

The distinction between what one is willing to express in public, and what one will express in private, highlights the importance of knowing people’s true attitudes for predicting their behavior. One technique for learning these attitudes is the bogus pipeline (Jones & Sigall, 1971). Imagine that you’re participating in a psychology study and the researcher asks you—in front of your peers, who you know are disgusted by the proliferation of reality shows—whether you like to watch The Voice and The Biggest Loser. Even if you actually enjoy settling in on the sofa and watching the trials and tribulations of people testing their singing voices or trying to lose weight, you might be reluctant to reveal this attitude in front of the other students. In fact, you might even be reluctant to admit to this guilty pleasure if it were just you and the experimenter in the room. But now imagine that the experimenter hooks you up to a polygraph machine—that is, a “lie-detector”—and asks you the same question. Now the experimenter seems to have a pipeline, or a direct route, to your true attitudes. Under these conditions, you might be much more likely to confess your enjoyment of these shows—better that than to be caught lying!

What makes the pipeline bogus, so to speak, is that often the experimenter isn’t attaching participants to a real mind-reading machine. What is important is that participants think that the experimenter knows exactly what they are thinking. And indeed, when people believe that there is a pipeline into their private thoughts, the attitudes they express predict their behavior better than the attitudes they express when concerned about self-presentation. The bogus pipeline is especially useful in situations where people are under strong pressures to express socially desirable attitudes, or to inhibit socially undesirable attitudes. For example, when child molesters are attached to what appears to be a lie detector, they are more likely to confess their sexual thoughts about children (Gannon et al., 2007).

Implicit Attitudes

Another approach to avoiding seeing merely what is socially desirable is to measure people’s implicit attitudes, which we introduced in chapter 3. These are attitudes people have but are not consciously aware that they have. Implicit attitudes predict some forms of behavior better than explicit attitudes—those that people can report consciously—because they are less influenced by self-presentational concerns about how they should and should not feel. Social psychologists have developed a number of techniques to measure implicit attitudes. For example, they often present participants with a word or an image (e.g., a picture of an African American face) and then observe how quickly the subjects recognize positive or negative words. If the word or image triggers negative feelings, participants should recognize negative words more quickly.

306

We will talk about this work in much greater detail when we discuss prejudice in chapter 10. But as you might imagine, it often is undesirable to admit to a prejudiced attitude toward African Americans, or Hispanics, or the physically handicapped. Yet many people clearly possess these prejudiced attitudes (Devine, 1989). Indeed, because people tend to underreport their level of prejudice on explicit measures, the correlation between implicit and explicit racial attitudes is weak (Dovidio et al., 2001). Because implicit attitudes lurk beneath people’s conscious awareness, they can predict subtle social behaviors that the people themselves do not recognize. For example, a person claims to have no prejudice toward African Americans, but when talking to African Americans, his body language tells a different story: He makes little eye contact, sits farther away, and interrupts the others’ sentences. These behaviors are poorly predicted by his explicit attitude (“I thought the conversation went great!”) but correlate highly with his implicit attitude (Dovidio et al., 2002).

Implicit attitudes are good predictors of many other types of behavior. For example, implicit attitudes toward smoking do a better job than explicit attitudes in predicting whether teens start smoking (Sherman et al., 2009), and implicit attitudes toward drinking better predict drinking behavior (Houben & Wiers, 2007). You may notice that both of these examples pertain to behaviors that are often engaged at the spur of the moment. And indeed, whereas explicit attitudes often do a better job of predicting more deliberate and reasoned behavior, implicit attitudes often fare better in predicting spontaneous behavior (Rydell & McConnell, 2006).

The Strength of the Attitude

Let’s return to our example of holding a negative attitude toward a certain movie while, at the same time, holding a positive attitude toward a romantic partner who wants to see the movie. How can we determine which attitude will win out, so to speak, and exert a greater influence on your behavior? In general, whichever attitude is stronger, or held more firmly, will exert a more potent influence. So if you really detest this kind of movie or the leading actor, and your liking for your partner is not especially strong, you might be more likely to bail on the movie outing. Stronger attitudes not only influence behavior more, they are also more enduring over time and resistant to change (Petty & Krosnick, 1995). So, in this example, your partner would have a very difficult time trying to persuade you to see this movie.

What determines the strength of an attitude? There are a number of factors to consider. First, an attitude tends to be stronger when it stems from a person’s own experiences rather than second-hand experiences they have heard about from others. Thus, if you see the trailer for a particular movie and think, “This stinks,” your attitude toward that movie is likely to be stronger than if you heard from an acquaintance that the movie is not even worth renting.

Because attitudes stemming from firsthand experience are stronger, they do a better job of predicting behavior. This effect was shown in a study that took advantage of a severe housing shortage at Cornell University during the 1970s (Regan & Fazio, 1977). In the throes of this shortage, many first-year students had to make do with temporary accommodations, such as sleeping on a cot in a dormitory lounge. Others were fortunate enough to be assigned permanent housing. Regan and Fazio contacted students from both groups and assessed their attitudes toward the housing crisis, asking them, for example, how much they had suffered and whether they thought the university was dealing with the situation effectively. Later, the experimenters gave participants the opportunity to take action, such as signing a petition or joining a committee of students to make recommendations about the crisis. Although the two groups of students had roughly equivalent attitudes toward the housing crisis, those who had to suffer through it personally (rather than simply hearing about it) had a much stronger connection between their attitudes and their behavior. The more negative their attitudes, the more likely they were to engage in behavior to try and correct the situation. For those without the personal experience, the connection between their attitudes and their behavior was much weaker.

307

Another factor that contributes to attitude strength is the extent to which the individual has a vested interest in the attitude. The greater our vested interest in an attitude, the stronger the attitude, and thus the better the attitude will predict behavior. Consider, for example, the current drinking age where you live. How bummed would you be if a law were passed that raised the drinking age by a couple of years? Would you do anything to prevent such a law from passing? Do you think that it might depend on how old you currently are? Sivacek and Crano (1982) studied just this situation in Michigan in the late 1970s, when a ballot was proposed to raise the minimum drinking age from 18 to 21. The researchers assessed students’ ages, their attitudes toward the proposal, and their interest in volunteering to campaign (e.g., call voters) against the proposal. Among students who were over 21, and thus had little vested interest in the drinking age issue, there was only a small relationship between their attitudes and their behavior. However, among those under 21, who would have been affected personally by the proposal, there was a much tighter relationship between their attitudes and behavior. The more against the proposal they were, the more likely they were to volunteer to work against it.

A final factor that influences attitude strength is the importance of the attitude domain. Two people may hold a similar attitude, but for one person the domain in question may be much more important. An avid sports fan and a casual fan may both like the Yankees. But sports are more important to the avid fan, who will have the stronger attitude about the Yankees and be more likely to purchase Yankees gear. Here’s a personal example from your current author. My wife and I, though both liking Thai food in general, dislike this local Thai restaurant in our area. But whereas you couldn’t drag me there with a pack of wild horses (I’d rather starve), she will concede to going there if others have already chosen that location. For me, food quality is more important, whereas for my wife, being socially amiable is more important. Our dislike is the same, but the importance of the domain (food quality vs. socializing) varies. When a domain is important to us, our attitude about it is more likely to influence our behavior. In a study of political attitudes and voting behavior, for example, Krosnick (1988) showed that the political policies that were most important to people exerted the strongest influence on their actual voting behavior.

The Accessibility of the Attitude

In situations where we have multiple attitudes that are relevant to a given behavior, the attitude that is most accessible will be the one most likely to guide behavior. We see this demonstrated in a classic study by Fazio and Williams (1986). They measured the accessibility of an attitude by measuring the speed (or latency) of a person’s response. The basic idea is that the more quickly a person indicates her attitude when asked, the more accessible that attitude is. Fazio and Williams examined whether more accessible attitudes toward the then American presidential candidates Ronald Reagan and Walter Mondale would predict people’s actual voting behavior. They assessed how quickly American citizens indicated their attitudes toward Reagan and Mondale. A few months later, after the election, they called those people and asked whom they voted for. What did the researchers find? The more quickly a participant indicated her attitude, the better that attitude predicted her voting behavior. For example, someone who more quickly said she had a positive attitude toward Reagan was more likely to vote for Reagan than someone who indicated the same positive attitude, but took longer to do so.

The fact that accessible attitudes predict behavior gives us important insights into why attitudes may at times fail to predict behavior. As we go about our daily affairs and interact with the world, we’re often focused on what is happening around us, and we’re not thinking of our attitudes. And unless we think about our attitudes, they are not accessible and thus have little influence on our behavior.

308

Can you recall a theory that we discussed in chapter 5 that explains what kinds of situations might make our attitudes accessible and thus more likely to influence our behavior? If you came up with self-awareness theory (Duval & Wicklund, 1972), you hit the nail on the head. To recap, this theory explains that people’s attention can be focused internally on the self or externally on the environment. When people are self-aware, they compare their current behavior with the standard, or attitude, that is relevant to their situation. If possible, they then adjust their behavior so that it matches that salient attitude.

Recall, for example, the study by Carver (1975) (see chapter 5) that initially measured participants’ attitudes toward the use of punishment in teaching. When participants were not self-aware, their preexisting attitude about punishment did a poor job of predicting the level of punishment they administered when trying to teach another person a vocabulary list. Yet when participants were made self-aware by looking in a mirror, their preexisting attitudes toward punishment were much better predictors of their willingness to open up a can of punishment. The interpretation here is that self-awareness made the preexisting attitude more influential.

But how is it that our attitudes actually guide our behavior? Do we just feel positively or negatively about something and then act accordingly? Or is there an intermediary step along the way? One influential theory proposes that our attitudes do not influence our behavior directly; rather, they do so through our intentions. According to the theory of planned behavior (Ajzen, 1985), attitudes are one of three ingredients in the intentions that we form (see FIGURE 8.14). The other two ingredients are subjective norms and perceived behavioral control. Attitude is your positive or negative evaluation of performing the behavior. Subjective norms are your perceptions of approval or disapproval of performing the behavior—that is, how much you perceive that others in your life think that a given behavior is a good or bad thing to do. And finally, perceived behavioral control is how much control you think you have over the behavior—whether you think you can do it or not. These three ingredients combine to shape your intentions, which in turn directly influence your behavior.

309

CONNECT ONLINE:

Check out our videos and additional resources located at: www.macmillanhighered.com/launchpad/greenberg1e