Efficiency of the Profile Likelihood Estimator for Case-Control Studies Under General Misspecification

• Activity Number: 330
• Type: Topic Contributed
• Date/Time: Tuesday, August 2, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Jennifer L. Wilcock*+ and Alan J. Lee
• Companies: University of Auckland and University of Auckland
• Keywords: case-control ; semiparametric efficiency ; asymptotic variance

Abstract

Scott and Wild (1997 and 2001) introduced a profile likelihood approach to estimation for generalised case-control studies and the method has recently been extended to accommodate multi-phase case-control designs (2010). In this talk we describe a simple and general approach which can be used to demonstrate the semiparametric efficiency of the Scott-Wild estimator in a variety of response-selective schemes whilst also allowing for a general misspecification of the model.

Our approach is based on projections and an adaptation of the method of Newey (1994). Previous work is extended in two ways. First, we treat the case of efficiency in a wide-sense where the estimation is carried out under possible general misspecification rather than under the assumption that the model is true. Second, our approach allows a variety of contexts to be handled by a single argument rather than each context requiring extensive specialised calculations as in Lee and Hirose (2008).