The evolution of Organizational Populations: Dynamic Resource Constraints and density-Depndent Competition among American Television Stations 1940-2000
The study concerns the emergence, evolution and organizational change in the population of American TV stations from 1940 - the date in which the very first commercial TV station appeared in the United States - to the year 2000. The empirical work is guided by hypotheses developed in the context of current ecological theories of organizations. Ecological theories differ from more conventional ways to think about organizations in that they: (i) Emphasize the role of selection rather than adaptation as the main engine of long-run change in the industrial and corporate world; (ii) Concentrate on populations of organizations, rather than on individual corporate entities, and (iii) Examine the life histories of all organizations in a given population observed over its entire history, rather than samples of different kinds of organizations followed for limited period of time. The main objectives of the study involves testing hypotheses about density-dependent organizational evolution, and exploring patterns of system dependence in the population of US television stations. The study is based on an event-oriented research design. In an event-oriented research design, sample units are observed to occupy only a finite (typically small) number of discrete states. The exact sojourn time in each state occupied during the observation period is recorded, and so is the timing of all transitions between states. Stochastic models are then estimated to understand factors that affect the transition rates. The empirical work involves the statistical analysis of the vital events experienced by each of the 1,348 commercial TV stations that ever existed in the United States during the period of observation. When individual organizational life histories are split into yearly spells, the size of the sample that we analyse consists of 34,861 valid observations. This split-spell design allows us to estimate the effects of time-changing covariates of theoretical relevance on rates of organizational founding, disbanding and change. Event history data maintain information on the natural timing of events. As such they can be easily aggregated to arrive at more conventional data structures such as, for example, time series of event counts. The results of the study promise to improve our understanding of the forces that regulate diversity in the organizational world in an institutional field ¿ the television sector ¿ in which issues of diversity are central and are likely to be of interest to a variety of parties including policy makers, the general public, and executives of media companies.