The Trait Approach: The Consistency Controversy

The Consistency Controversy

Whether they endorse the Big Five dimensions or not, trait theorists agree that individ-uals’ personalities can be described in terms of stable and enduring traits. After all, when we say that someone is friendly and warm, we are doing more than describing how he acted on a particular occasion. Instead, we are describing the person and, with that, providing some expectations about how he will act on other occasions, in other settings. But is this right? Is someone’s behavior stable in this way?

HOW CONSISTENTARE PEOPLE ?

In one classic study, researchers examined the behavior of schoolchildren—and, in par-ticular, the likelihood that each child would be dishonest in one setting or another (Hartshorne & May, 1928). Quite remarkably, the researchers found little consistency in children’s behavior: Children who were inclined to cheat on a school test were often quite honest in other settings (e.g., an athletic contest), and vice versa. Based on these findings, it would be misleading to describe these children with trait labels like “honest” or “dishonest”—sometimes they were one, and sometimes the other.

Some 40 years ago, Walter Mischel reviewed this and related studies, and concluded that people behave much less consistently than a trait conception would predict, a state of affairs which has been referred to as the personality paradox (Mischel, 1968). Thus, for example, the correlation between honesty measured in one setting and honesty measured in another situation was .30, which Mischel argued was quite low. Mischel noted that behaviors were similarly inconsistent for many other traits, such as aggres-sion, dependency, rigidity, and reactions to authority. Measures for any of these, taken in one situation, typically do not correlate more than .30 with measures of the same traits taken in another situation. Indeed, in some studies, there is no detectable corre-lation at all (Mischel, 1968; Nisbett, 1980). These findings led Mischel to conclude that trait conceptions of personality dramatically overstate the real consistency of a person’s behavior.

WHY AREN ’T PEOPLE MORE CONSISTENT ?

How should we think about these results? One option is to argue that our personalities are, in fact, relatively stable just as the trait approach suggests, but acknowledge that situations often do shape our behavior. Given a red light, most drivers stop; given a green light, most go—regardless of whether they are friendly or unfriendly, stingy or generous, dominant or submissive. Social roles likewise often define what people do independent of their personalities. To predict how someone will act in a courtroom, for example, there is little point in asking whether he is sociable, careless with money, or good to his mother. What we really want to know is the role that he will play—judge, prosecutor, defense attorney, or defendant.

We reviewed studies indicating that the influence of a situation can be incredibly powerful—leading ordinary college students to take on roles in which they are vicious and hurtful to their peers. It’s no wonder, then, that there is sometimes lit-tle correspondence between our traits and our behavior and less consistency in our behavior than the trait perspective might imply. The reason, in brief, lies in what’s called the power of the situation. Because of that power, our behavior often depends more on thesetting we are in than on who we are.

Sometimes, though, our behavior does depend on who we are. Particularly in weak situations—ones in which the environment provides few guides for our behavior—our personalities shape our actions (Figure 15.4). Even in strong situations—ones in which the environment provides clear guides for our behavior—different people react to the situation in somewhat different ways, so that their behavior in the end reflects the inter-action of the situation with their personality (Fleeson, 2004; Magnusson & Endler,1977). Moreover, it’s not a matter of chance how a particular person reacts to this situation or that one; instead, people seem to be relatively consistent in how they act in certain types of situations. Thus, for example, someone might be punctual in profes-sional settings, but regularly late for social occasions; they might be shy in larger groups, but quite outgoing when they are with just a few friends.

Evidence for these points comes from many sources, including a study in which chil-dren in a summer camp were observed in a variety of situations—settings, for example, in which they were teased or provoked by a peer, or settings in which they were approached in a friendly way by a peer, or settings in which they were scolded by an adult (Cervone & Shoda, 1999; Mischel, Shoda, & Mendoza-Denton, 2002). In this study, the researchers relied on behavioral data—data based on observations of specific actions— and these data showed that each child’s behavior varied from one situation to the next. For example, one child was not at all aggressive when provoked by a friend, but responded aggressively when scolded by an adult. Another child showed the reverse pat-tern. Thus, the trait label aggressive would not consistently fit either child—sometimes they were aggressive and sometimes they were not.

There was, however, a clear pattern to the children’s behavior, but the pattern emerges only when we consider both the person and the situation. As the investigators described it, the data suggested that each of the children had a reliable “if . . . then . . .” profile: “If in this setting, then act in this fashion; if in that setting, then act in that fashion” (Mischel et al., 2002). Because of these “if . . . then . . .” patterns, the children were, in fact, reason-ably consistent in how they acted, but their behaviors were “tuned” to the situations they found themselves in. Thus, we need to be careful when we describe any of these children as being “friendly” or “aggressive” or “helpful,” relying only on global trait labels. To give an accurate description, we need to be more specific, saying things like “tends to be friendly in this sort of setting,” “tends to be helpful in that sort of setting,” and so on.

ARE SOME PEOPLE MORE INFLUENCED BY THE SITUATION THAN OTHERS ?

There is one more complexity we must keep in mind as we consider how personality and situations interact to shape behavior. Some individuals are more consistent than others across situations, or, turning this around, some individuals are more flexible than others. This difference among people is assessed by the Self-Monitoring Scale, developed by Mark Snyder and designed to assess the degree to which people are sen-sitive to their surroundings and likely to adjust their behaviors to fit in. The scale includes items such as “In different situations and with different people, I often act like very different persons.”

High self-monitors care a great deal about how they appear to others, and so, at a cocktail party, they are charming and sophisticated; in a street basketball game, they “trash talk.” In contrast, low self-monitors are less interested in how they appear to

others. They are who they are regardless of the momentary situation, making their behavior much more consistent across situations (Figure 15.5; Gangestad & Snyder, 2000; M. Snyder, 1987, 1995). This suggests that the extent to which situations deter-mine an individual’s behavior varies by person, with situations being more important determinants of high self-monitors’ behavior than of low self-monitors’ behavior.

How consistent individuals are also varies at the cultural level of analysis. Americans, for example, are relatively consistent in how they describe themselves, no matter whether they happen at the time to be sitting alone, next to an authority figure, or in a large group (Kanagawa, Cross, & Markus, 2001). By contrast, Japanese partici-pants’ self-descriptions varied considerably across contexts, and they were far more self-critical when sitting next to an authority figure than when they were by themselves. There also cultural differences in how consistent individuals want to be. In one study, researchers asked American and Polish participants how they would respond to a request to take a survey about beverage preferences. When asked to imagine they had previously agreed to such requests, American participants said they would again agree to the request—apparently putting a high value on self-consistency. Polish partici-pants, by contrast, were much less concerned with self-consistency, and so were less influenced by imagining that they had agreed to similar requests in the past (Cialdini, Wosinka, Barrett, Butner, & Gornik-Durose, 1999).