Determining causal relationships in leadership research using Machine Learning: The powerful synergy of experiments and data science

Machine Learning (ML) techniques offer exciting new avenues for leadership research. In this paper we discuss how ML techniques can be used to inform predictive and causal models of leadership effects and clarify why both types of model are important for leadership research. We propose combining ML and experimental designs to draw causal inferences by introducing a recently developed technique to isolate heterogeneous treatment effects. We provide a step-by-step guide on how to design studies that combine field experiments with the application of ML to establish causal relationships with maximal predictive power. Drawing on examples in the leadership literature, we illustrate how the suggested approach can be applied to examine the impact of, for example, leadership behavior on follower outcomes. We also discuss how ML can be used to advance leadership research from theoretical, methodological and practical perspectives and consider limitations.

Automated summary

This paper discusses how machine learning techniques can be used in leadership research, including the use of ML techniques to inform predictive and causal models of leadership effects, as well as combining ML with experimental designs for causal inference. The paper also provides a step-by-step guide on designing studies that combine field experiments with ML to establish causal relationships with maximal predictive power.