Academic journal article Management Revue

Techniques of Event History Modeling. New Approaches to Causal Analysis, 2nd Edition

Academic journal article Management Revue

Techniques of Event History Modeling. New Approaches to Causal Analysis, 2nd Edition

Article excerpt

Blossfeld, Hans-Peter / Rohwer, Götz: Techniques of Event History Modeling. New Approaches to Causal Analysis, 2nd Edition Lawrence Erlbaum Associates, Mahwah/NJ and London 2002, Paperback, 310 pp, $ 36.00, Hardcover: $ 79.95

Event history analysis has sometimes been called a new causal approach for the social sciences. It differs from traditional approaches in several ways which are roughly encapsulated by the terms event and history. Firstly, event history data are far more informative than cross-sectional data or even panel data. Whereas panel data capture states of a process at pre-determined survey points, event history data inform about the course of events between the survey points as well, i.e. they inform about the history of the process under study. secondly, event history models analyze events. An event marks the transition from one discrete state to another. Considering turnover as an example, one might think of quitting as an event. Quitting marks the transition from organizational membership to a state of no organizational membership. Statistically speaking event history analysis allows for the explicit modeling of time. In event history models "time [...] normally serves as a proxy variable for a latent causal factor that is difficult to measure directly" (p. 228). For example, human capital theory assumes that firm specific human capital is accumulated with length of tenure. In this respect time (length of tenure) serves as a proxy variable for human capital specificity. Drawing again on the turnover example, length of tenure should be negatively related to employee turnover. Event history models then calculate the "instantaneous probability" of an event (quit) at a certain point of time (tenure), given that the event has not occurred before that point of time. This is reflected by the transition or hazard rate, which serves as the "dependent variable" in event history models.

Event history models have been successfully applied to various research topics of Organizational Behavior (OB). Besides turnover, researchers have studied absence periods using event history models, the life cycle of organizations, mobility between unemployment and employment, full-time and part-time work or different types of employment, as for example employment and self-employment. Blossfeld and Rohwer provide lots of further examples and references in the introduction to Techniques of Event History Modeling. New Approaches to Causal Analysis. However, this is not a book on OB. It is rather a comprehensive reader on longitudinal data analysis which offers certain points of reference for researchers from different disciplines who are interested in the dynamics of social processes. The book serves three goals: First, the authors "demonstrate that event history models are an extremely useful approach to uncover causal relationships or to map out a system of causal relations" (p. vii). second, they introduce the computer program Transition Data Analysis (TDA) by Rohwer and Pötter (2002) which is widely recognized as the most flexible and complete program for analyzing event histories. Third, the book serves as a supplement and update to the first edition (Blossfeld and Rohwer 1995) and to the textbook by Blossfeld et al. (1989). This review provides a brief overview of the book focusing on the first two goals with regard to research on OB.

The introductory chapter of Techniques of Event History Modeling gives an overview of typical observation plans. The authors discuss cross-sectional data, panel data and event history data and the (dis-)advantages of these different designs. In the next section they develop a view on causality, which leads over to the basic statistical concepts of event history analysis. As event history analysis follows a causal and probabilistic approach, it offers a very natural way to understand and interpret the indeterminacy of human behavior. This is elaborated by the transition rate concept.

The second chapter deals with the structure of event history data. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.