اعتماد به نفس کاذب و پنج عامل بزرگ

It has been consistently observed that people are generally overconfident when assessing their performance. In the current study, participants completed Goldberg's Big Five personality inventory and then completed a cognitive task designed to assess overconfidence (defined as the difference between confidence and accuracy). Extraversion significantly predicted overconfidence (with the other Big Five factors controlled statistically). In addition, openness/intellectance significantly predicted confidence and accuracy but not overconfidence (again, with the other Big Five factors controlled statistically). Theoretical implications and implications for future research are discussed.

There is growing interest in individual differences in judgment and decision making. The major developments in this area have been in relating intelligence to judgment strategies, largely with the intent to resolve the competing claims of normative models (e.g., Stanovich & West, 2000). However, there are other ways in which judgment and decision making may be affected by individual differences. One example is overconfidence, which is only negligibly correlated with intelligence (Stanovich and West, 1998 and Stanovich and West, 1999). Overconfidence is a judgmental error in which people overestimate their own accuracy. Specifically, overconfidence is defined as a positive difference between confidence and accuracy. Participants typically answer many general knowledge questions in a two-alternative, forced choice format, and then rate their confidence in each answer on a half-range scale, from 50 to 100%. Three primary statistics are computed: average confidence, average accuracy, and overconfidence, which is the difference between average confidence and average accuracy. A very large number of studies have reported positive average overconfidence values (for one review, see Budescu, Wallsten, & Au, 1997). That is, people are generally more confident than they are accurate. It is easy to conceive domains wherein overconfidence will be detrimental in personal, professional, financial, sporting, and other domains. It is therefore of interest to explore what kinds of people are more overconfident than others. There is an ongoing debate as to the utility of individual differences in predicting overconfidence and other biases. Some research has failed to uncover such relationships (Stankov & Crawford, 1997; Wright & Phillips, 1979), whereas other research has found relationships (e.g., Pallier et al., 2002). For example, several non-clinical variants of pathological traits predict overconfidence. Participants with narcissistic personalities have been found to be more overconfident than non-narcissists. They are also just as willing to bet on their more-inflated confidence as others, resulting in significantly greater losses in a laboratory-based betting game (Campbell & Goodie, 2003, Experiment 1). Narcissists also are more likely to display the related phenomenon of overclaiming (Paulhus, Harms, Bruce, & Lysy, 2003). That is, narcissists are more likely to claim to know the meaning of terms that do not actually exist. Indeed, narcissists' behavior is evidence of self-enhancement biases found across a range of social psychological studies (e.g., Colvin & Block, 1994). In contrast to narcissism, subclinical depression is also linked to lower levels of overconfidence—even underconfidence—in general knowledge tests (Stone, Dodrill, & Johnson, 2001). However, depressed individuals may report greater overconfidence when judging real-life events (Dunning & Story, 1991). Other personality traits have also been linked to overconfidence. For example, research on confidence (with accuracy controlled statistically) has revealed a negative association between negative affectivity and confidence, and positive associations between optimism, need for cognition, self-monitoring, and self-efficacy (Wolfe & Grosch, 1990; but see Olivares, 1993, for the finding of no link between overconfidence and self-monitoring). Likewise, researchers have uncovered a positive association between authoritarian personality and overconfidence (Olivares, 1993). Furthermore, no association has been found between openness to experience (from the NEO-PI) and overconfidence (r=.00; Olivares, 1993). The association between openness and accuracy also failed to reach significance (r=.11, p=.11). Finally, research has uncovered a positive association between two variables associated with extraversion [proactiveness from the True Self-Report Inventory ( Irvine, 1999)] and the activity facet of extraversion from the NEO-PI-R ( Costa & McCrae, 1992) and overconfidence as measured by a principle components analyses of responses to a series of intellectual tests ( Pallier et al., 2002). The association between overall extraversion and overconfidence was .15, but did not reach significance at the p<.01 level selected by the researchers. In sum, there is a diverse range of associations between various personality variables and overconfidence. The goal of the present research is to advance this area of research by examining the link between the most commonly used, high-level measure of personality, the Big Five, and overconfidence. We assess overconfidence with a broad general knowledge test in order to minimize systematic biases caused by preexisting knowledge (Lichtenstein & Fischhoff, 1977). The Big Five approach to personality posits that individual differences that are socially relevant and useful—in other words, personality traits—are encoded on common language (see Goldberg, 1993; John & Srivastava, 2002). The analysis of these trait terms has led to the discovery of five basic factors to describe personality: extraversion (Factor I), agreeableness (Factor II), conscientiousness (Factor III), neuroticism (Factor IV), and openness to experience/intellectance (Factor V). We predict that extraversion will be positively associated with overconfidence based on the following. First, there is some evidence that facets of extraversion are linked to overconfidence (Pallier et al., 2002), and the lack of an association with overall extraversion may reflect a lack of power (r was .15). Second, extraversion is linked roughly r=.40 to narcissism (e.g., Paulhus & Williams, 2002) and narcissism correlates with overconfidence. Third, extraversion is linked to dispositional optimism ( Williams, 1992), which is suggestive of increased overconfidence. Finally, there is evidence that traits linked to positive affectivity, such as extraversion, are also linked to confidence when accuracy is controlled statistically ( Wolfe & Grosch, 1990). Agreeableness contains such descriptors as “sympathetic” and “pleasant.” It is difficult to see how such traits should be related to overconfidence. Likewise, agreeableness correlates negatively with narcissism (Paulhus & Williams, 2002). We thus do not make any firm predictions regarding agreeableness. Likewise, conscientiousness contains such descriptors as “deliberate.” Conscientiousness may predict global performance in settings that require diligence, which might be reflected in greater accuracy scores, but may not be expected to predict overconfidence. In contrast, neuroticism is predicted to be negatively associated with overconfidence. This is consistent with findings that dispositional optimism is negatively associated with neuroticism (Wolfe & Grosch, 1990), that negative affectivity in some forms is negatively related to confidence (with accuracy controlled statistically; Williams, 1992), and that depression (another correlate of neuroticism) is associated with underconfidence (Stone et al., 2001). Finally, we predict that openness to experience/intellectance will be associated with accuracy but not overconfidence. We make this prediction even though past research failed to yield a significant association between openness as measured by the NEO (Olivares, 1993) and accuracy. The lack of statistical significance in this study may have been the result of the selection of 25 difficult questions for the outcome variable. The exclusive use of difficult questions likely restricted the variance in the dependent measure and thus resulted in a highly conservative test of the hypothesis. Intellectance appears to be related to both intelligence and academic achievement (Helson, 1985; McRae & Costa, 1985). In turn, good confidence calibration has been related to both intelligence and academic achievement (Hacker, Bol, Horgan, & Rakow, 2000).