Leadership Process Models: A Review and Synthesis

Abstract

In organizational research, studying "processes" is important for uncovering and understanding the underlying causal mechanisms in a predictor-mediator-outcome logic. Processes answer "how" and "why" questions and provide more complete explanations about phenomena. Our focus in this review is on studies of leadership processes, which we systematically analyze to report on the state-of-the science. In doing so, we present a two-dimensional target-centric taxonomy to integrate previous research: The taxonomy distinguishes the target's level (i.e., individual follower, team, organizational, and extra-organizational) as well as the type of leadership processes that affect either the target's development or leverage of resources. Our review indicates that the predominantly studied leadership "meta" process model looks at the effect of leader traits or behaviors on performance-related outcomes through cognitive, affective, or behavioral leveraging factors. This "meta" model points to several important and understudied processes including a leader's influence on the target's development or work context. We also identify two largely overlooked yet critical issues for leadership process research: Modeling the role of time and that of multiple processes through which leadership effects manifest themselves in organizations. Using our taxonomy, we provide several reflection points that can guide the development of genuine and thoughtful leadership process theories. We conclude by urging future leadership process research to embrace multi-process, multi-level, and time-sensitive models.

Leader–member exchange (LMX) differentiation and work outcomes: Conceptual clarification and critical review

“…Moreover, it implies a reciprocal causal relationship between LMX quality and LMX differentiation processes: The initial development of LMX quality impacts LMX differentiation, and subsequent changes in LMX differentiation affect changes in LMX quality. Time‐sensitive designs would be helpful in exploring these temporal processes (see Fischer et al, , for a more detailed discussion of temporality and leadership) and would assist in measuring the actual process of differentiating how the relationships with different members change over time and what “events” may occur to influence such differentiation. Such “events” could include (in addition to changes in team membership, as described above) violations of trust, contract breaches as perceived by the leader, or delegation of unmet responsibilities.…”

On doing relevant and rigorous experiments: Review and recommendations

“…In order to correct for this bias, the authors should examine whether the mediator is actually endogenous (via a Hausman test; see Hausman, 1978) and use an instrumental variable estimator (Antonakis et al, 2010; Emsley et al, 2010) instead of the standard OLS or ANOVA. Finally, researchers should also mind the problem of specious mediation (modeling the wrong mediator that correlates with the true mediator), as well as causally unidentified mediation chains (Fischer et al, 2016): A causally identified mediation model needs to have at least a unique exogenous predictor per endogenous mediator, the unique exogenous variable/s must be excluded from the y equation, and the system of equations must be estimated using an instrumental variable estimator (Antonakis et al, 2010; Antonakis et al, 2014b).…”

“…The logic is clear: The level of an outcome today might be dependent on yesterday's level and the researcher might be interested in explicitly modeling the adjustment dynamics. However, the lagged variable might be endogenous itself (Bollen and Curran, 2006; Fischer et al, 2016; Wooldridge, 2010). Where is the issue?…”

“…Concerning the experimental research in the other fields, we selected a random sample of articles of non‐OM disciplines: A random selection of articles should allow to have an unbiased representation of a discipline's average experimental study and maintain a feasible workload level for the coders (cf. Antonakis et al, 2010; Bergh and Perry, 2006; Fischer et al, 2016; Kacmar and Whitfield, 2000). Using a search term in the “topic” field on Web of Science, we downloaded the list of all articles explicitly mentioning the words “experiment” or “experiments” either in the title, in the abstract or in the keywords, randomly ordered them, and manually selected the first 50 articles that were relevant for our analysis.…”

Underveis Menns og kvinners lederkarrierer

“…Kunnskap om hvordan menns og kvinners karrierer etableres og utvikles er likevel begrenset, og i ledelsesforskningen understrekes behovet for bedre innsikt i prosessen og betydningen av tidsdimensjonen i analysene (Fischer, Dietz & Antonakis, 2016). Nyere bidrag har ved bruk av registerdata identifisert faktorer som kan svekke hennes sjanser for å ha lederstilling i sammenligning med hans (Hardoy, Schøne & Østbakken, 2017a, 2017b.…”