Explained Variance for Bias of Item Discrimination Parameter, Item Difficulty Parameter, and ICC Recovery for the Empirical Histogram Method.

Note. Mean-centered aggregated values (i.e., bias in each condition) were used as data. Therefore, residual variance is reduced. Absolute effect sizes η2 should be interpreted as the respective factor’s relative contributions to the total amount of bias. To avoid overfit, models only included main effects and two-way interactions.Explained Variance for Bias of Item Discrimination Parameter, Item Difficulty Parameter, and ICC Recovery for the Empirical Histogram Method.