Testing for Publication Bias in Political Science

If the publication decisions of journals are a function of thestatistical significance of research findings, the publishedliterature may suffer from "publication bias." This paper describesa method for detecting publication bias. We point out that toachieve statistical significance, the effect size must be largerin small samples. If publications tend to be biased againststatistically insignificant results, we should observe thatthe effect size diminishes as sample sizes increase. This propositionis tested and confirmed using the experimental literature onvoter mobilization.