Discrete Survival Time Models: An Application on Marriage Duration

Abstract

In survival analysis which is used in the social and physical sciences, it is usually assumed that the observed process is continuous. Since this assumption is not appropriate for most of the survival time data structure, survival times are measured wrongly and unreliable results are obtained for the discrete survival time data. Continuous time survival models used for the time data have represented the structure of data in the studies regarding health sciences. The usage of the discrete time survival models in social sciences is more common since the structure of the studied data is more appropriate for the discrete time models. In this study, discrete time survival models are examined theoretically and were applied to "Research on Domestic Violence against Women in Turkey, 2008" data received from Turkish Statistical Institute. In order to examine the factor effecting the duration of marriage, discrete time survival models have been used and achieved results have been interpreted.

Keywords

• complementary log-log model
• cox regression
• discrete time survival models
• logit model
• non-proportional hazards