Latent Class Analysis and its Uses

Posted by

Latent Class Analysis (LCA) is a type of model-based cluster analysis that has numerous advantages over traditional cluster analysis techniques such as hierarchical cluster analysis. LCA provides a flexible and powerful approach to categorical data analysis (McCutcheon and Hagenaars, 1997).

LCA has been demonstrated to be a valuable tool for assessing measurement error, identifying flawed questionnaire items, assessing mode effects and adjusting for nonresponse bias. Furthermore, as LCA is a cluster analysis technique, it is extremely useful for carrying out person-centred (as opposed to variable-centred) analysis. Integrating person-centred and variable-centred approaches could provide a more complete understanding of data and further assist policy makers with their decision-making.

Quality Improvement Funding was obtained to promote the use of LCA across the GSS in order to produce better quality and more relevant outputs for users of Official Statistics.

This power point presentation provides an introduction to LCA and describes its advantages over other types of cluster analysis techniques. It also described the various types of LCA and examples of applications and potential applications of this technique.

This guide briefly describes LCA and how it can be used. It also provides an applied example to illustrate how to carry out and interpret an LCA using R (R Foundation for Statistical Computing, 2011) and SAS1.

Debbie Cooper
Office for National Statistics



1 SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc