Table 11.

Academic effect sizes with and without left‐side trim and fill imputation.

Outcome	N of effect sizes	Estimated N of missing effect sizes (SE of N)	Hedges’ g (95% CI) without Imputation	Hedges’ g (95% CI) with L0 trim and fill imputation
Composite of academic outcomes	113	21 (6.99)	0.23 (0.17, 0.29)	0.14 (0.06, 0.22)
General academic ability	24	0 (2.99)	0.21 (0.06, 0.36)	NA
Language/literacy	46	11 (4.42)	0.24 (0.16, 0.33)	0.13 (0.03, 0.24)
Mathematics	36	0 (3.64)	0.28 (015, 0.42)	NA

Note: Outcomes with an NA indicate that there was enough symmetry that the trim and fill algorithm did not estimate any effect sizes to be missing and, therefore, no effect sizes were imputed. Random‐effects effect size estimates were for imputed and nonimputed estimates to facilitate comparison; therefore, these effect size estimates may differ from the cluster‐robust estimates presented elsewhere. Social studies and science were not included in this analysis because they had less than 10 effect sizes. L0 means left–side imputation.

Abbreviation: CI, confidence interval.