English Français |
Philippe Bocquier, UCLouvain
Ashira Menashe-Oren, Universite catholique de Louvain (UCL)
Carren Ginsburg, University of the Witwatersrand, Johannesburg
This paper gives a proof of concept by combining two previously-separated life history methods, sequence analysis and event history analysis, introducing time-varying life histories as a covariate in time-to-failure analysis. Using Agincourt Health and Demographic Surveillance Data, we compare sequence analysis and other characterisation of migration itineraries (2nd order and manual grouping) to evaluate the effectiveness of these different procedures. As a side product, the paper also proposes a methodology to extend sequence analysis to the whole population, thus lifting the constraint of sub-sampling.
Presented in Session 120. New perspectives in demographic surveys