Neutralizing substitutes for leadership theory: leadership effects and common-source bias.

The purpose of this research was to examine alternative models of substitutes for leadership theory given the general lack of empirical support for the moderating effects postulated by the theory. On this basis, the research posited that the effects of substitutes also could be conceptualized as mediated relations. The research examined moderated and mediated relations for several sets of leader behaviors and substitutes that have been examined in the literature. The research design sampled 49 organizations, with 940 subordinates rating 156 leaders. Results, although generally not supportive of the moderator or mediator hypotheses, essentially demonstrated that leadership matters. The findings also suggest that prior significant effects in substitutes literature may be merely a statistical artifact, resulting from common-source bias.