Abstract:
|
Marginal model is a popular approach for longitudinal data. For count responses with preponderance of zeros, marginal zero-inflated models may be applied. In this talk, we develop a test for zero-inflation, or more generally, zero modification, for marginal Poisson regression models for longitudinal data. Without specification of zero-inflated Poisson models, we develop our test based on the generalized estimating equation (GEE) methodology. The performance of the test in terms of type I errors and power are evaluated by simulation studies. The results show that the test statistic performs well under a wide range of conditions. The test is illustrated with a real data example
|