An exploratory factor analysis using maximum likelihood rotation provided a five-factor solution which supported our theoretical typology of five leadership types (see Table 1). This solution is similar to that found by Pearce et al. (2003). As Tabachnick and Fidell (2001) suggested, factor scores were estimated through the regression approach rather than averaging the items. Factor loadings were used as item weights to create factor scores. This approach was recommended in order to cope with multicollinearity issue (Basilevsky, 1994). These five factor scores were used as variables for further analysis. The leadership variables were aversive, directive, transactional, transformational, and empowering