Using Mixed-Effect Growth Models to Examine Time as a Predictor of Interest and Between-Firm Differences Over Time

Abstract

Panel data, where observations of entities are repeated over time, are common in strategic management research. However, explorations of the role of time on predictors of interest are often unexplored. In this chapter, we illustrate how the use of mixed-effect growth models can enhance theory and research in strategic management by exploring changes in outcomes of interest over time. Mixed-effects models allow for testing both within and between effects, while also calculating specific intercepts (firm average values) and slopes (trajectories of specific firms over time) using empirical Bayes estimates. We also illustrate how a discontinuous growth model could be used to assess differences in firm intercepts and slopes surrounding exogenous events (e.g., global pandemics) without requiring a control group.

Citation

Schepker, D.J. and Bliese, P.D. (2023), "Using Mixed-Effect Growth Models to Examine Time as a Predictor of Interest and Between-Firm Differences Over Time", Hill, A.D., McKenny, A.F., O'Kane, P. and Paroutis, S. (Ed.) Methods to Improve Our Field (Research Methodology in Strategy and Management, Vol. 14), Emerald Publishing Limited, Leeds, pp. 79-99. https://doi.org/10.1108/S1479-838720220000014005