Mathematical Statistics.

Abstract

In genetic association studies outcome dependent sampling is often used in order to increase power.When analyzing the data, correction for the ascertainment scheme generally has to be made to avoid bias.Such correction is however not available in standard statistical methods when the data structure and/or the ascertainment scheme is complex.In this report three simulation based approaches that can be used for correction of known ascertainment schemes are described.These methods provide parameter estimates and are flexible in terms of what statistical models and ascertainment schemes can be handled.Some simulations are conducted to evaluate the methods.