404 - 
Effects of education, occupational characteristics and cohort on 
the "family cycle" 
Andreas Diekmann 
1. Introduction 
Patterns of family formation and dissolution are largely in- 
fluenced by socio-economic factors like education and income 
of spouses. A theoretical framework explaining the relations 
between these variables and the timing and duration of mar- 
riage is given by Becker's (1981) theory of family econo¬ 
mics. 
Many multivariate empirical studies dealing with the sociolo- 
gical and economic causes of divorce and age at marriage 
either use individual dichotomous response data (e.g. the 
attribute never married at some age) or aggregate data (e.g. 
divorce rates in different states). In the former case data 
are analysed by logit- or probit-models or some other dicho¬ 
tomous-response model (see e.g. Becker, Landes, Michael, 
1977; Galler, 1979; Hogan, 1978; McDonald and Rindfuss, 
1981; Michael, 1979; Waite and Spitze, 1981) while in the 
latter case conventional regression methods can be applied 
(Freiden, 1975; Preston and Richards, 1975; Sander, 1985). A 
third line of empirical research has recently begun to uti- 
lize event-history data on the "familiy cycle". Age at entry 
into marriage or marriage duration are regarded as "arrival 
times" which are analyzed by appropriate methods of survival 
analysis. This kind of analysis is usually superior because 
it builds on very informative event-history data, yields a 
model for the process, and allows for many derivations of 
characteristics of the process (see Hannan, Tuma, Groene¬ 
veld, 1977; Teachman, 1982; Heaton, Albrecht, Martin, 1985; 
Tuma and Michael, 1985; Sørensen and Sprensen, 1985; Diek- 
mann and Mitter 1984a; Diekmann, 1986a). 
In this paper marriage and divorce data from the cumulative 
General Social Survey 1980 to 1984 on about 9000 West German 
households are analyzed by techniques of event history ana¬ 
lysis. Three aspects are stressed in this article: To pre-