event-history analysis

A form of longitudinal study in which the unit of enquiry is not the individual or social group but the socially significant event, for example, change of occupation, spell of unemployment, change of marital status, or other significant life-event . Arguably, the technique is of long standing, having an obvious precursor in the life-history method of the Chicago School. However, a number of recent conceptual and methodological advances have created great interest in the approach, and there is now an extensive literature (especially on life-history and work-history analyses).

Conceptual advances include the concept of life-course analysis, which focuses attention on the overlaps between an individual's experiences and those coincident in historical time, for example changes in government policy, the experiences of other family members, or status of significant other individuals (see, Transitions: The Family and the Life Course in Historical Perspective, 1978).

At the level of research techniques, advances in dealing with problems of attrition and recall have extended the reliability of panel data, and the development of a number of statistical techniques and packages for handling recurrent binary events have made the analysis of events (and their associated ‘time windows’) both more sophisticated and easier to present (see, Analysis of Panel Data, 1986). The Cox proportional hazard model is a good example of one such technique.

The papers collected in, Life and Work History Analyses (1991) give a good idea of the issues and recent developments in the field. One of the leading exponents of this form of analysis, Hans-Peter Blossfeld, has also produced two (co-authored) texts-Event-History Analysis (1989) and Techniques of Event-History Modelling (1995)-which give an excellent account of the explanatory potential of event-history techniques for causal modelling of longitudinal data. See also sequence analysis.