Revisiting the mediating role of leader–member exchange in transformational leadership: the differential impact model

Abstract

Transformational leadership (TFL) has been proposed as an essential antecedent of leader–member exchange (LMX), which in turn affects outcomes in organizations. We extend this mediation hypothesis in two ways by proposing a differential impact model, which we test on three organizational outcomes: employee job satisfaction, employee organizational commitment, and leader effectiveness. First, we extend LMX’s mediational impact—which has previously only been tested for employee outcomes—to leader effectiveness. Second, we argue that this mediation will be stronger for outcomes that are more proximal rather than distal to dyadic relations between leader and followers (high proximity: job satisfaction; medium proximity: organizational commitment; low proximity: leader effectiveness). Meta-analytic structural equation modelling based on 132 studies revealed that LMX mediates TFL’s relationships with employee outcomes (more strongly for job satisfaction than for commitment), but not with leader effectiveness, whereas TFL showed a stronger direct link to leader effectiveness. The findings suggest that TLF and LMX contribute differentially to organizational outcomes depending on their proximity to dyadic relations between leaders and followers. The differential impact model uncovers leadership effectiveness processes, integrates influential leadership theories, and highlights the importance of distinguishing between different outcome measures and the processes facilitating them.

Keywords:

• transformational leadership
• leader-member exchange
• leader effectiveness
• meta-analysis

Acknowledgements

We would like to thank the EnBW—Energie Baden Würtenberg AG and the WDN—WISE Demographic Network for financial support of this study. Furthermore, we would like to thank Mike WL Cheung for his valuable comments on the application of MASEM. Additionally, we gratefully acknowledge the suggestions of Annebel de Hoogh (University of Amsterdam) and the invaluable help with coding from Daniela Gutermann, Kyra Nitsche, Miriam Summ, and Vanessa Grebe.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplemental data and research materials

Supplemental data for this article can be accessed at http://dx.doi.org/10.1080/1359432X.2016.1170007.

Notes

1. We argue that the distinction between an individual and a group focus in TFL might especially underscore the meaningfulness of the relational aspects driving the outcomes of transformational leadership, and thereby matches our focus on LMX as a potential mediator. However, in the current study, we do not differentiate between the individual and group-focused aspects of TFL, because (a) the purpose of this study is to more generally investigate the complementarities of leader behaviours and relations for employee and leader outcomes, and, most importantly, (b) this differentiation has only recently emerged, which means that the number of studies which could be used for a differentiated analysis is rather low and would severely limit the statistical power. We strongly encourage researchers to investigate the LMX-mediational model for individual-focused and group-focused TFL separately in future research (also see our discussion).

2. Moderator analyses revealed that the reliability of leadership measures did not affect the magnitude of the effect sizes (see Table 4). Regarding leadership measures, we found that effect sizes were stronger when the MLQ was used compared to other TFL measures for the association between TFL and effectiveness (.46 vs. .29), as well as job satisfaction (.44 vs. .37). Study design had significant effects on the associations between TFL and outcomes, as well as on the inter-correlations between the outcome measures. More precisely, cross-sectional designs showed higher effect sizes than longitudinal designs regarding the correlation between TFL and effectiveness (.46 vs. .07), TFL and job satisfaction (.47 vs. .22), TFL and commitment (.39 vs. .26), and between effectiveness and job satisfaction (.47 vs. .12), and effectiveness and commitment (.32 vs. .15). Interestingly, the reverse held regarding the association between job satisfaction and commitment, with the effect sizes being higher in longitudinal designs (r = .61) compared to cross-sectional designs (r = .50).

3. Student samples reported higher correlations than samples from business contexts regarding the association between TFL and effectiveness (.60 vs. .39) and between job satisfaction and commitment (.65 vs. .52). The cultural contexts affected six effect sizes; however, the direction of the effects varied. Studies conducted in the USA contexts reported stronger correlations between LMX and job satisfaction (.48 vs. .42), TFL and satisfaction (.50 vs. .35), between effectiveness and job satisfaction (.64 vs. .28) and commitment (.33 vs. .16) compared to studies conducted outside the USA, whereas studies conducted outside the USA reported stronger correlations between TFL and commitment (.42 vs. .29) and between job satisfaction and commitment (.58 vs. .45) compared to studies conducted in the USA contexts.

4. Older samples of leaders and followers reported lower associations between TFL and job satisfaction, and leaders’ older age was associated with lower correlations between leader effectiveness and job satisfaction (see Table 4). With regard to gender distributions, a higher percentage of male followers was associated with lower correlations between TFL and commitment.

Additional information

Funding

We would like to thank the EnBW—Energie Baden Würtenberg AG and the WDN—WISE Demographic Network for financial support of this study.