Affordable Access

Publisher Website

Population-Based Case-Control Studies-Chapter 38

Elsevier Science & Technology
DOI: 10.1016/s0169-7161(09)00238-7
  • Design


Publisher Summary This chapter discusses the design and analysis of case–control surveys conducted with a view to fitting binary regression models. Case–control data is special because the response (outcome) variable is the most important design variable and because the selection probabilities typically differ enormously, often by several orders of magnitude. Varying selection probabilities can distort the mean structure if not taken into account and estimates produced by standard programs may be inconsistent. The use of different strata for cases and controls is an example of this type. The inefficiency of standard design weighting for case–control data should not be unexpected. It is well known that weighting in general tends to be inefficient when the weights are highly variable. In case–control studies, the variation in weights is about as extreme as it can get and no experienced survey sampler would be surprised to find that weighting is not very efficient under these circumstances.

There are no comments yet on this publication. Be the first to share your thoughts.


Seen <100 times