Population Selection Bias: The Case of Density Dependence

Social science often focuses on organizations, institutions, and phenomena that have become socially or economically important. This can create a population selection bias. Even if one samples all relevant organizations, and not only successful ones, one may selectively sample populations of organizations. The effects of such a selection bias are illustrated by using the case of density dependence in organizational entry and exit rates. While several studies have found non-monotonic density dependence in entry and exit rates, most studies only examine and estimate models of vital rates in populations that have become large. This creates a population selection bias that can give rise to spurious non-monotonic density dependence. The implications of this bias are discussed for organizational ecology and more generally for the selection of the populations and empirical settings studied in social science.