Content area
Full Text
The purpose of this study was to examine whether self-awareness of leaders (define as agreement between self and other leadership ratings) would moderate the validities of : (a) common selection devices (e.g., tests, experience) to predict leader behavior of U.S. Naval Academy students (n = 91), and (b) leader behavior to predict performance of naval officers (n = 158). Leaders were classified into over-estimation, under-estimation, or agreement categories, based on the difference between their self-ratings of leadership and the ratings provided by superiors and/or subordinates. Results indicated that the magnitudes of the correlations between predictors and leader behavior, as well as between leader behavior and performance, varied as a function of agreement category. Also, the correlations between leader behavior and performance were highest for officers in the agreement category. The primary implication of this study is that self-awareness should be considered in attempts to predict leader behavior and performance. Additional methodological and practical implications are discussed.
Self-appraisals have been investigated from a number of perspectives. These include focusing on agreement between self-ratings and ratings from other sources, biases in self-ratings, and the effects of rater accuracy upon performance. Three general conclusions have resulted from this work. First, self-ratings tend to be inflated, suffering from leniency and social desirability biases (e.g., Podsakoff & Organ, 1986). Second, self-ratings are less highly related to ratings by others (i.e., peers, supervisors, or subordinates) than peers', supervisors', and subordinates' ratings are with one another (Harris & Schaubroeck, 1988). Relatedly, self-ratings are also less accurate than ratings from peers or supervisors, when compared to objective criterion measures (e.g., Hough, Keyes, & Dunnette, 1983). Third, inaccurate self-raters (i.e., those with self-ratings that differ greatly from observer ratings) are poorer performers than their more accurate counterparts (e.g., Bass & Yammarino, 1991; Flocco, 1969).
SELF-RATINGS AND INDIVIDUAL CHARACTERISTICS
These findings suggest that self-ratings are unreliable indicators of behavior. However, self-ratings may be useful if they are thought of as providing information about the self-rater's dispositions. A number of studies have suggested that individuals who rate themselves in specific ways have particular individual characteristics. For example, Farh and Dobbins (1989) discovered that leniency bias in self-ratings was related to self-esteem. Compatible with consistency theory (e.g., Korman, 1970), and contrary to self-enhancement theory (e.g., Greenwald, 1980),...