We used preregistered multilevel logistic regression analyses to predict (decision) inconsistencies and found a significant interaction effect of journal_type*time in the expected direction: a steeper decline in (decision) inconsistencies in statcheck journals than matched controls.

Several journals started implementing #statcheck in their peer review process. Their hope is that requiring a “clean” statcheck report will significantly reduce statistical reporting inconsistencies in the articles they publish. BUT! This remains an empirical question.