Comparing methods of classifying life courses: sequence analysis and latent class analysis

Authors

  • Yu Han NiDi (the Netherlands Interdisciplinary Demographic Institute) University Groningen
  • Aart C. Liefbroer NiDi (the Netherlands Interdisciplinary Demographic Institute) University Medical Centre Groningen Vrije Universiteit Amsterdam
  • Cees H. Elzinga NiDi (the Netherlands Interdisciplinary Demographic Institute) Vrije Universiteit Amsterdam

DOI:

https://doi.org/10.14301/llcs.v8i4.409

Keywords:

life course, sequence analysis, latent class analysis

Abstract

We compare life course typology solutions generated by sequence analysis (SA) and latent class analysis (LCA). First, we construct an analytic protocol to arrive at typology solutions for both methodologies and present methods to compare the empirical quality of alternative typologies. We apply this protocol to develop and compare SA- and LCA-derived family-life typologies for women born between 1960 and 1964 in 15 European countries, using data from the Family and Fertility Survey. This paper contributes to the use of these classification techniques in four different ways. First, we present guidelines on how to establish the number of classes or clusters to use. Second, we show how to evaluate the stability of these clusters. Third, we provide a way to evaluate the validity of these clusters and finally, we provide for a formal heuristic to relate the stochastically defined latent classes to the distance-based clusters found with SA.


Downloads

Published

2017-10-26