Assessing and relaxing assumptions in quasi-simplex models

The quasi-simplex model makes use of at least three repeated measures of the same variable to estimate its reliability. The model has rather strict assumptions about how various parameters in the model are related to each other. Previous studies have outlined how several of the assumptions of the quasi-simplex model may be relaxed using more than 3 waves of data. It is unclear however whether the assumptions of the quasi-simplex model are overly strict. In other words, it is not known whether relaxing the assumptions results in better models or different substantive conclusions with regard to the reliability of survey measures. Using data from the British Household Panel Survey this paper shows how the assumptions of the quasi-simplex model can be relaxed. We conclude that relaxing the assumptions in practice seldom leads to a better model or different conclusions than the traditional quasi-simplex model.