LOGISTIC-REGRESSION MODELS FOR BINARY PANEL-DATA WITH ATTRITION

We discuss ways of analysing panel data when the response is binary an d there is attrition or drop-out. In general, informative or non-ignor able drop-out models are non-identifiable and arbitrary constraints on the drop-out model must be imposed before carrying out a statistical analysis. The problem is particularly acute when predictors as well as response variables are lost by attrition. We describe a likelihood-ba sed method for dealing with the drop-out process in this difficult cas e and show how the effect of non-identifiability can be reduced by imp orting additional data from a cross-sectional survey of the same popul ation. The methods are primarily motivated by data from the 1987-92 Br itish Election Panel Study and the 1992 British Election Study.