"Accounting for Endogeneity in Associations between HR Systems, Leadership, and Employee Attitudes"

Abstract

The purpose of this research was to make stronger causal inferences about the associations between HR systems, leadership, and employee attitudes in an observational field sample. We controlled for a number of potential sources of endogeneity with longitudinal employee survey data that was collected over a period of eight years. We first conducted propensity score matching to reduce the likelihood that selection or sorting effects influenced the results. We then compared lagged panel regression models that tested if HR systems and/or leadership likely had causal effects employee attitudes, if omitted third variables were a potential source of endogeneity, or if reverse causality (employee attitudes cause perceptions of HR systems or leadership) could explain the observed effects. The results showed that there was a reciprocal relationship between HR systems and person-organization fit perceptions, leadership caused perceptions of the organization’s employment reputation, and that person-organization fit caused perceptions of leadership. We discuss how causal inference affects theory development and provide suggestions about how scholars can improve research designs to make stronger causal inferences with observational field data.