Title: Publication bias, and a model for sensitivity analysis

Abstract:

Publication bias is a serious problem in the interpretation of reviews of the published literature. We will describe a relatively simple model for sensitivity analysis, based on empirical evidence that research studies giving significant results are more likely to be published. Some applications will be discussed.