Abstract
The purpose of this article was to determine proximal and distal antecedents of leadership in the U.S. Army’s Officer Candidate School (OCS). A model composed of motivation to lead, leadership self-efficacy, implicit leadership, organizational commitment, general cognitive ability, and personality was proposed. Results from a longitudinal examination of 1,232 officer candidates suggest partial support for the model, and limited evidence for enlisted experience as a moderator. Candidate personality (partially mediated by interest in leadership and leadership self-efficacy) best predicted leadership performance during OCS and peer ratings of leadership potential. Implications for OCS selection and models of leadership performance are discussed.
Notes
1 Defined by the authors as “A desire to behave in an active and assertive manner involving activities that in this society are often viewed as predominately masculine and thus to meet managerial role requirements” ( CitationButler, Lardent, & Miner, 1983, p. 497).
2 Note that we use the term “personality” differently, in some ways more broadly and in some ways more narrowly, in the present research than is typically the case. First, the framework used here is performance-specific, and does not include aspects of personality that are thought to be unrelated to performance in a military context. In other words, this framework does not capture the full spectrum of “personality” as it is typically defined. Second, this framework includes scales, such as motivation for physical activity, that are not typically considered part of an individual’s “personality.” Despite these differences, we use the term “personality” throughout the report as shorthand for this framework.
3 Interrater reliability estimates were computed using G(q, k; CitationPutka, Le, McCloy, & Diaz, 2008), in which k is equal to the harmonic mean of the number of raters per ratee (avg. k = 5.30). The corresponding single-rater coefficients (i.e., G[q, 1]) were .50 and .47 for the company-grade and field-grade ratings, respectively.
4 Regression weights are unstandardized.
5 The mean and standard deviation on the AFQT for this sample was 84.86 and 13.45, respectively. This is higher and more range-restricted than the population standard deviation for college graduates as determined by the CitationBureau of Labor Statistics’ (2005) National Longitudinal Survey of youths (M = 78.17, SD = 17.62). For example, if we were to correct the correlations between AFQT and leadership performance and leadership potential for range restriction using Lawley’s formula (cf. CitationHunter & Schmidt, 1990), the magnitude of the correlation with leadership performance would be .12 rather than .09. It may also be that AFQT, which was developed as an indicator of general cognitive ability for the population of American youth between 18 and 23 years of age, is simply not as good of a measure of general cognitive ability for the college graduate population because of a lack of sufficient discrimination at the upper end of the score distribution.
6 However, it should be noted that the relationship between cognitive ability and leadership in this research was much smaller in this study than meta-analytic investigations would suggest ( CitationJudge et al., 2004).