Application areas within the social sciences include:

  • Births, Deaths, Marriage formations and dissolutions (Demography/Sociology)
  • Morbidity and hospitalizations (Public Health and Medicine)
  • Arrests and re-arrests/recidivism) (Criminology)
  • Unemployment durations, Job changes/promotions (Economics)
  • Mergers and bankruptcies (Business and Marketing)
  • Residence changes (Urban planning)
  • Entry into and exit from political power (Political Science)
  • Reliability of products (Technology)
  • etc…

Common in the above-listed areas is that they deal with time until occurrence of some event of interest (birth, death, divorce, arrest, employment, bankruptcy, product-failure, etc). Depending on the area of application, data of these types are also known as survival data, failure-time data, duration data, event-history data, or generally time-to-event data.   

Goals in the analysis of such data include describing the distributional shape of the time-to-event, comparing survival experiences of different groups, and modelling the association between background characteristics and some measures of survival. Such a measure may reflect quantum of the event of interest (the proportion experiencing the event) or its tempo (the speed at which it occurs).

Methodological issues of interest to the research group include (both in the frequentist and Bayesian framework):

  • Competing Risks (multiple causes of failure)
  • Clustering and multi-level modelling (frailty)
  • Selection bias and multi-process modelling
  • Time varying effects and dynamic modelling
  • Adjusting for anticipatory covariates
  • etc…