12
The Assessment of Implicit and Explicit Self-Esteem: Lessons from Motive Research
Coming from outside of this research area, we have been struck by the almost universal suspicion that social and personality psychologists have recently voiced regarding self-report measures of self-esteem. The measures have been criticized as too global, too static, too culture-bound, too easily conflated with narcissism, too easily obscured by self-presentation and not really important to effective functioning. Other self-related personality variables, such as self-regulation, self-discrepancies, or self-determination have been put forward as alternative constructs that are more important to healthy functioning. Many researchers interested in self-esteem have shifted their attention to implicit measures such as the Implicit Association Test (Greenwald & Banaji, 1995) or the Initials Preference Task (Bosson, Swann, & Pennebaker, 2000). This field-wide consensus questioning the value of the traditional measurement of self-esteem led us to assume that there was probably little value in measuring people’s self-esteem with self-report scales.
It was with surprise then that we found ourselves impressed when reading over the 10 items of the widely used Rosenberg’s (1965) self-esteem scale. The items have an essential face validity that is disarming. They clearly capture the feelings of self-acceptance and self-worth that are central to a humanistic view of self-esteem (Rogers, 1951). Items such as “I feel that I am a person of worth” and “On the whole I am satisfied with myself” certainly seem to reflect the sentiments of someone with high self-esteem. One cannot but imagine that a parent or teacher would consider it important that their child or student reported positive responses to these items. Alternatively, a parent or teacher would be concerned if a child endorsed one of’ the reverse-scored items such as “I feel I do not have much to be proud of.”
The psychometric qualities of the Rosenberg self-esteem scale (SES) also seem excellent, despite it being brief and easy to administer (Blascovich & Tomaka, 1991). The scale is uni-factorial, with high internal consistency and good test-retest reliability. Its validity has been supported by correlations with peer reports and with other self-esteem related measures (Blascovich & Tomaka, 1991). Rosenberg self-esteem appears to function well as a marker of positive mental health. It is positively related to other well-being indicators such as life satisfaction and positive and negative affect, and it is negatively related to indicators of psychological distress (Schimmack & Diener, 2003). It has been shown to be unrelated to demographic factors such as gender, age, and marital status, and it is only moderately related to social desirability (Fleming & Courtney, 1984).
Why then is there such angst about self-report measures of self-esteem? One clear problem is that recent research has shown that some pretty disturbed individuals with rather serious personality problems, such as antisocial behavior and narcissism, often report very high levels of self-esteem (Bushman & Baumeister, 1998). Another problem is that there is evidence that people with high self-esteem sometimes behave in more maladaptive ways than those with low self-esteem. For example, after receiving ego-threatening negative feedback, people with high self-esteem have been shown to make risky and unwise choices about what level of challenge they can handle (Baumeister, Heatherton, & Tice, 1993). A third problem is that research has shown that self-esteem may serve a primarily defensive function rather than reflecting deep and stable feelings of self-worth. For example, there is evidence that when faced with mortality salience many people will inflate their estimates of self-esteem (Greenberg, Solomon, & Pyszczynski, 1997).
Despite these problems, one wonders why researchers have not concluded that self-reported self-esteem is a generally useful indicator of a persons subjective sense of worth and value but that it may be vulnerable to distortion in the face of extreme personality pathology or extreme situations. Such a conclusion would lead to the practical solution that in order to get a full picture of an individuals functioning, one would want to collect additional information—e.g., rule out narcissism, recent failure or mortality experiences, ensure stability over time, etc. Researchers have tended not to choose the route of supplementing self-report self-esteem measures with such ancillary measures, instead they have shifted their attention to nonreactive or implicit measures of self-esteem.
Implicit self-esteem has been defined as an automatic, over-learned, and nonconsious evaluation of the self that guides spontaneous reactions to self-relevant stimuli (Greenwald & Banaji, 1995). The most commonly used method to assess implicit self-esteem is the Implicit Association Test (IAT) (Greenwald & Farnham, 2000), which is a computerized categorization task that measures the ease with which an individual automatically associates pleasant and unpleasant words with the self (Greenwald, McGhee, & Schwartz, 1998). A tendency to quickly associate pleasant rather than unpleasant words with the self is thought to reflect high implicit self-esteem. The IAT self-esteem measure demonstrated superior reliability, stability, and predictive validity relative to other implicit measures of self-esteem (Bosson et al., 2000). Interestingly, the IAT was moderately positively correlated with the Rosenberg scale, and it mirrored the Rosenberg in predicting how raters would evaluate an essay writers self-esteem and self-competence, whereas other implicit measures were unrelated to the Rosenberg and failed to predict raters’ evaluations (Bosson et al., 2000).
The IAT self-esteem thus seems to be the most promising of the implicit self-esteem measures on psychometric grounds (there is ongoing debate regarding the IAT, cf. Karpinsld, 2004), but there does seem to be general enthusiasm for response time measures of implicit self-esteem (Fazio & Olson, 2003). IAT self-esteem is marginally positively related to the Rosenberg self-esteem scale, and similarly predicts certain outcomes, but the two measures share only a small percentage of variance. The question that naturally arises then is how researchers should use these two relatively distinct measures of self-esteem. We would like to suggest that the history of work on implicit and explicit motives may serve as a useful guide for work on implicit and explicit self-esteem (Bosson et al, 2000). In the following section, we outline some insights derived from 50 years of motivation research on how to build a theory of implicit and explicit self-esteem.
Three distinct motives have received considerable attention over the previous 50 years: the need for achievement (“a recurrent concern with standards of excellence”), the need for power (“a recurrent concern with having impact or influence on others”), and the need for intimacy (“a recurrent concern with establishing warm, reciprocal relations with others”). From the earliest work in this area, researchers developed both implicit and self-report measures of each motive. Implicit measures were based on coding the thematic content of imaginative stories that participants told in response to ambiguous picture cues presenting achievement or social scenes. The coding systems were derived by arousing a particular motive (e.g., a convivial party for need for intimacy) and then comparing the stories written under aroused conditions with those written under neutral conditions (e.g., a classroom setting). The coding systems were objective and high levels of inter-rater reliability could be attained after a brief training period.
The early motive researchers did not expect the implicit and self-report measures of motives to be related to each other, or to predict similar outcomes. Indeed, McClelland (1951) conceptualized personality as consisting of three major systems, traits, motives, and schemas, and he believed that self-report measures of motives actually belonged in the category of a schema. That is, the self-report motive measures reflected how people thought about themselves rather than how they naturally and spontaneously organized their thoughts in relation to goals and incentives. A critical early study demonstrated that the implicit and explicit measures of achievement motivation were uncorrelated and predicted entirely different sets of outcomes (deCharms, Morrison, Reitman, & McClelland, 1955). Explicit achievement motivation was associated with being influenced by expert opinions whereas implicit achievement was associated with better performance on tasks that involved internalized standards of excellence. Many subsequent studies have confirmed that explicit and implicit motives predict different outcomes (Koestner, Weinberger, & McClelland, 1991; Woike, Gershkovich, Piorkowski, & Polo, 1999).
Despite the early evidence that implicit and explicit motives were independent and could both contribute to explaining behavior, motivation research became dominated by questions regarding the psychometric credentials of the implicit measures, repeated attempts to design valid self-report measures, and arguments about which type of measure displayed greater predictive validity.
This long-running implicit versus explicit debate was finally “resolved” in 1989 when McClelland restated his earlier view that it should be possible to build a more comprehensive theory of behavior by including both implicit and explicit measures of the three motives (McClelland, Koestner, & Weinberger, 1989). The following five central points were made:
Implicit and explicit motives represent different levels of personality and their measures will typically be uncorrelated with each other.
Implicit and explicit motives will each importantly influence behavior (and knowing about both of them is better than knowing about only one of them).
Implicit and explicit motives will predict different kinds of outcomes with implicit motives predicting spontaneous, self-initiated patterns of behavior over time whereas explicit motives predicting response tendencies in particular social situations.
Implicit and explicit motives are responsive to distinct sets of environmental stimuli (i.e., activity incentives such as challenge level for the implicit motive and social incentives such as whether a task is described as achievement-relevant for the explicit motive).
Implicit and explicit motives have distinct developmental histories with the key formative era for implicit motives being the first two years of life and the key era for explicit motives being middle childhood.
How is this history of implicit and explicit motives relevant to current work on self-esteem? We think there are three parallels for self-esteem research that have already become evident. First, one could have predicted that the implicit measures would come under attack because their psychometric qualities pale in comparison to those of self-report scales. (David McClelland was always dubious of the high reliability of self-report scales which he felt were based on asking redundant questions to participants who felt compelled to respond consistently.) Second, one could have predicted that many researchers would become discouraged by the fact that implicit and explicit measures were uncorrelated and predicted different outcomes. (The absence of a positive correlation between self-report and implicit measures of the achievement motive was central to Entwisle’s (1972) influential critique of motive research.) Third, it could have been predicted that researchers would respond to this situation by spawning additional implicit and explicit measures.
What would someone schooled in the motive literature advise regarding future work on implicit and explicit self-esteem? The following four key points stand out:
Researchers should settle on the implicit measure that seems to have the best psychometric credentials (but recognize that no implicit measure will ever attain the same levels of reliability as explicit self-report scales).
Researchers should be untroubled by the lack of relation between implicit and explicit measures.
Researchers should work to identify the class of outcomes that are uniquely associated with implicit and explicit self-esteem.
Researchers should outline the distinctive situational and developmental factors that impact on implicit and explicit self-esteem.
Above all, self-esteem researchers would do well to avoid becoming mired in a debate regarding which type of measure is better. Instead, they should celebrate the fact that self-esteem can be assessed at both an implicit and explicit level and be inspired to build a complex, multilevel theory of self-esteem that can take its place amidst other recent comprehensive, multilevel conceptualizations of personality (McAdams, 1995, 2001; Sheldon, 2004).
References
Baumeister, R. E, Heatherton, T. E, & Tice, D. M. (1993). When ego threats lead to self-regulation failure: Negative consequences of high self-esteem. Journal of Personality & Social Psychology, 64, 141–156.
Blascovich, J., & Tomaka, J. (1991). Measures of self esteem. In J. L. Robinson, P. R. Shaver, & L. S. Wrightsman (Eds.), Measures of personality and social psychological attitudes (Vol. 1, pp. 115–160). New York: Academic Press.
Bosson, J. K., Swann, W. B. Jr., & Pennebaker, J. W. (2000). Stalking the perfect measure of implicit self-esteem: The blind men and the elephant revisited? Journal of Personality & Social Psychology, 79, 631–643.
Bushman, B. J., & Baumeister, R. E (1998). Threatened egotism, narcissism, self-esteem, and direct and displaced aggression: Does self-love or self-hate lead to violence? Journal of Personality ó-Social Psychology, 75, 219–229.
deCharms, R., Morrison, H. W., Reitman, W. R., & McClelland, D. C. (1955). Behavioral correlates of directly and indirectly measured achievement motivation. In D. C. McClelland (Ed.), Studies in motivation (pp. 414–423). New York: Appleton-Century-Crofts.
Entwisle, D. R. (1972). To dispel fantasies about fantasy-based measures of achievement motivation. Psychological Bulletin, 77, 377–391.
Fazio, R. H., & Olson, M. A. (2003). Implicit measures in social cognition research. Their meaning and uses. Annual Review of Psychology, 54, 297–327.
Fleming, J. S., & Courtney, B. E. (1984). The dimensionality of self-esteem. Journal of Personality and Social Psychology, 39, 921–929.
Greenberg, J., Solomon, S., & Pyszczynski, T. (1997). Terror management theory of self-esteem and cultural worldviews: Empirical assessments and conceptual refinements. Advances in Experimental Social Psychology, 29, 61–136.
Greenwald, A. G., & Banaji, M. R.(1995). Implicit social cognition: Attitudes, self-esteem, and stereotypes. Psychological Review, 102, 4–27.
Greenwald, A. G., & Farnham, S. D. (2000). Using the implicit association test to measure self-esteem and self-concept. Journal of Personality and Social Psychology, 79, 1022–1038.
Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74, 1464–1480.
Karpinski, A. (2004). Measuring self-esteem using the Implicit Association Test: The role of the other. Personality and Social Psychology Bulletin, 30, 22–34.
Koestner, R., Weinberger, J., & McClelland, D. C. (1991). Task-intrinsic and social extrinsic sources of arousal for motives assessed in fantasy and self-report. Journal of Personality, 59, 57–82.
McAdams, Dan P. (1995). What do we know when we know a person? Journal of Personality, 63, 365–396.
McAdams, Dan P. (2001). The person: An integrated introduction to personality psychology. New York: Harcourt-Brace.
McClelland, D. C. (1951). Personality. New York: Sloane.
McClelland, D. C., Koestner, R., & Weinberger, J. (1989). How do self-attributed and implicit motives differ? Psychological Review, 96, 690–702.
Rogers, C. (1951). Client-centered therapy. Boston: Houghton-Mifflin.
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.
Schimmaek, U., & Diener, E. (2003). Predictive validity of explicit and implicit self-esteem for subjective well being. Journal of Research in Personality, 37, 100–106.
Sheldon, K. M. (2004). Optimal human being: An integrated multi-level perspective. Mahwah, NJ: Laurence Erlbaum Associates.
Woike, B., Gershkovich, I., Piorkowski, R., & Polo, M. (1999). The role of motives in the content and structure of autobiographical memory. Journal of Personality & Social Psychology, 76, 600–612.