Subject specific and population average models for binary longitudinal data: a tutorial
Keywords:Autocorrelation, British Household Panel Survey, hierarchical models, logistic regression, marginal models, mixed effects models, random effects models, repeated-measures analysis
AbstractUsing data from the British Household Panel Survey, we illustrate how longitudinal repeated measures of binary outcomes are analysed using population average and subject specific logistic regression models. We show how the autocorrelation found in longitudinal data is accounted for by both approaches, and why, in contrast to linear models for continuous outcomes, the parameters of population average and subject specific models for binary outcomes are different. To illustrate these points, we fit different models to our data set using both approaches, and compare and contrast the results obtained. Finally, we use our example to provide some guidance on how to choose between the two approaches.
Authors who published with Longitudinal and Life Course Studies Volumes 1–9 agreed to the following terms:
1. Authors retain copyright and grant the Journal right of first publication with the work, simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
2. Following first publication in this Journal, Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal, provided always that no charge is made for its use.
3. Authors are permitted and encouraged to post their work online (e.g. in institutional repositories or on their own website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work.