|
Abstract:
|
A recently designed study of the impact of oral contraceptive use on the risk of bone fracture employed the randomized recruitment scheme of Weinberg and Wacholder (1991). Data from such schemes may be analyzed within the two-phase framework. A potential complication in the present study is that the differential response rates between cases and controls may result in participation bias. Viewing participation as an additional stage in a multi-stage sampling scheme, weighted likelihood estimation provides a obvious approach. In this work we seek to extend the profile-likelihood approach or profile pseudo-likelihood to improve statistical efficiency. Specifically, we propose to introduce additional weights to account for the participation mechanism. Since the fracture study is on-going, we illustrate the methods using birth weight data from North Carolina.
|
