Fig 5. P-value and replication outcomes.

The relationship between p-value category and replication is robust to a meta-analysis, with a pooled effect of 0.49. There is no evidence for heterogeneity between studies (I² = 41%, 95% CI [0%, 80%], τ² = 0.0205, 95% CI [0, 0.49] and Q (3) = 5.06, p = 0.17), however with only 4 studies included in the meta-analysis there may be low power to detect heterogeneity. For all studies, the effect is in the same direction.