Abstract

We describe a standard, easily interpretable effect size estimate for parapsychological research, the proportion index (pi or ?). This estimate gives the hit rate on a scale on which the null value is al ways .50, no matter what the number of equally likely alternatives. The index ? is helpful in the design, analysis, and meta-analysis of parapsychological research. First, it can be used to help select the best choice of the number of response alternatives. Second, signifi cance tests and confidence limits for a can be readily computed within a study. Of particular importance for meta-analytic purposes, tests of heterogeneity of a set of ?s and of contrasts among them are described as well as a coefficient of the robustness of replication of a set of ?s.