OBJECTIVES: The aims of the study were to determine whether differences in leadership self-perception/behaviour in healthcare researchers may influence research performance and to evaluate whether certain leadership characteristics are associated with enhanced leadership efficiency in terms of motivation, effectiveness and satisfaction. DESIGN AND PARTICIPANTS: All Faculty of Medicine Professors at Imperial College London (n=215) were sent the Multifactor Leadership Questionnaire (MLQ) Self form as a means of evaluating self-perception of leadership behaviours. MAIN OUTCOME MEASURES: For each professor, we extracted objective research performance measures (total number of publications, total number of citations and h index) from 1 January 2007 to 31 December 2009. The MLQ measured three leadership outcomes, which included motivation, effectiveness and satisfaction. Regression analysis was used to determine associations. RESULTS: A total number of 90 responses were received, which equated to a 42% response rate. There were no significant correlations between transformational, transactional or passive/avoidant leadership behaviours and any of the research performance measures. The five transformational leadership behaviours (ie, idealised attributes (IA), idealised behaviours (IB), inspirational motivation (IM), intellectual stimulation (IS), individual consideration (IC)) were highly significant predictors of leadership outcomes, extra effort (all B>0.404, SE=0.093-0.146, p<0.001), effectiveness (IA, IM, IS, IC B>0.359, SE=0.093-0.146, p<0.001; IB B=0.233, SE=0.103, p=0.026) and satisfaction (IA, IM, IS, IC B>0.483, SE=0.086-0.139, p<0.001; IB B=0.296, SE=0.101, p=0.004). Similarly, contingent reward was a significant predictor of extra effort (B=0.400, SE=0.123, p=0.002), effectiveness (B=0.353, SE=0.113, p=0.002) and satisfaction (B=0.326, SE=0.114, p=0.005). CONCLUSIONS: This study demonstrates that transformational leadership and contingent reward positively influence leadership efficiency in healthcare researchers. Although we did not show an association between leadership behaviours and research performance metrics, further studies using contextual performance measures at team and organisational levels are required.