Abstract. Publication bias hampers the estimation of true effect sizes. Specifically, effect sizes are systematically overestimated when studies report only significant results. In this paper we show how this overestimation depends on the true effect ...