ABSTRACT
This chapter focuses on the ubiquity and extent of heterogeneity in behavioral research. It reviews three classes of meta-analytic techniques. The chapter also reviews meta-analytic techniques that assess and adjust for publication bias. It discusses how meta-analytic techniques can accommodate study-level covariates. The Many Labs and registered replication reports approach allows data to be integrated via meta-analysis—with the data from each laboratory treated as an independent replication study—to provide an assessment of heterogeneity. There are a number of meta-analytic techniques based on standardized effect sizes on standardized scales such as the J. Cohen’s d scale and the correlation scale. The chapter discusses two techniques—so-called single paper meta-analysis and multilevel multivariate meta-analysis—based on basic summary information such as means, standard deviations, and sample sizes. It is often of interest in meta-analysis to assess the relationship between the effect of interest and one or more study-level covariates.
