A SEMIPARAMETRIC RANDOM EFFECTS APPROACH TO UNOBSERVED HETEROGENEITY IN COUNT DATA MODELS OF ORGANIZATIONAL FOUNDING

The paper discusses a semiparametric random-effects approach to the pr oblem of unobserved population heterogeneity in organizational researc h based on models for pooled cross-sectional time series count data. T he analytical value of this approach rests in its ability to produce e stimates of the structural parameters that do not depend on any specif ic assumption about the distribution of the heterogeneity components i n the population. The practical value of the method proposed is illust rated in an empirical application to processes of organizational found ing, and to the relation between density dependence and unobserved het erogeneity in spatially distributed organizational populations. The em pirical evidence produced suggests that future studies of organization al founding at the population level will have to account for variation in observed as well as unmeasured (or unobservable) variables.