Many Labs 2 (psychology)
To download only this data file: ManyLabs2.rds (52 KB)
To download all BEAR datasets, click here.
Many Labs 2 (psychology)
Reference: Klein et al. (2018).
Snapshot date: Jan 2026
Research question: replications of 28 classic and contemporary effects in psychology, with the same effects tested across many samples and research sites.
Data availability: the authors make their replication package available through OSF. We use the paper-included Many Labs 2 analyses from the key table and the site-level aggregated replication results from the public data package.
Data processing: we retained the 28 preregistered analyses included in the paper. Each row is a site-level replication estimate for one analysis, so the dataset has many more rows than analyses. We use the Many Labs 2 analysis identifier as metaid. Thus, a single finding such as Kay.1 contributes one BEAR row for each sample/site estimate, rather than one row for each alternative statistical scale reported for the same sample. The additional grouping variable distinguishes analyses designated as primary or secondary by the original authors.
The retained effect-size measure is the correlation scale. We used the reported r estimate as b, its reported variance to compute se where available, and then computed z = b / se. For two analyses where the variance of r was not available but a signed equivalent test statistic was available, we used that statistic as z and derived the implied standard error. Original-study effects were matched by analysis where available and stored separately from the replication estimates.
Additional grouping variable: primary or secondary analysis designation.
Model of z-values
The fitted mixture model is shown over the empirical distribution of absolute z-values. The solid line is a mixture of half-normals, with selection. The dashed line shows the distribution without selection. If there are inequalities (e.g. studies reporting p < 0.05) the histogram resamples values from the appropriate set.
