Content area
Full Text
The Breadth-based Adjective Rating Task (BART) operates as an indirect measure of self-esteem by quantifying the extent to which people describe expectancy consistent self-information at a higher level of abstraction than expectancy inconsistent information about the self. A series of three studies demonstrated evidence for the validity of the BART as a measure of global, trait-like aspects of self-esteem. The BART correlated positively with explicit measures of self-esteem, negatively with self-esteem instability and self-reported depression, but did not correlate with implicit or indirect measures of self-esteem. In Study 4, the test-retest reliability of the BART was adequate, albeit low, in two independent samples. These findings provide preliminary support for the BART as a promising indirect paper-and-pencil measure of self-esteem.
The issue of how to best capture and assess attitudes and self-esteem has been a dilemma for social and clinical psychologists alike. Traditional explicit attitude measures rely on self-report and are susceptible to a number of biases, particularly for sensitive content areas such as self-esteem where individuals may be unwilling or unable to provide accurate reports. In an attempt to resolve this problem, many social psychologists have turned to implicit measures of attitudes. An implicit attitude measure assesses constructs that reside outside of conscious awareness and control (Greenwald & Banaji, 1995; Fazio & Olson, 2003). Many implicit measures of self-esteem have been proposed (Karpinski & Steinman, 2006; Greenwald & Farnham, 2000; Herts, Sakuma, & Pelham, 1999; Spalding & Hardin, 1999; Kitayama & Karasawa, 1997); however, concerns remain about their reliability, validity, and interpretation (see Bosson, Swann, & Pennebaker, 2000; Fazio & Olson, 2003). In a comprehensive study of seven implicit self-esteem measures, Bosson et al. (2000) found that only two measures displayed adequate psychometric properties: the self-other Implicit Association Test (IAT; Greenwald & Farnham, 2000) and the initial preference task (Nuttin, 1985, 1987; Kitayama & Karasawa, 1997). These two measures displayed modest levels of test-retest reliability and at least some ability to predict criterion variables relevant to self-esteem.
However, there are lingering concerns about both of these self-esteem measures. The self-other IAT provides a measure of the positive and negative associations a person has with the self compared with others. Thus, the self-other IAT is not a pure measure of self-associations, but rather a joint...