Table 1. Meta-analyses of Effectiveness of Randomized Clinical Trials Assessing School Anti-Bullying Interventions.

Outcome Variable	Duration of intervention, mean (95% CI), wk	Length of follow-up, mean (range), wk	k	No. of participants		Meta-analysisa		Heterogeneityb		Publication biasc
				Intervention groups	Control groups	Cohen d, mean (95% CI)	FDR corrected P value	Q statistic P value	I2 statistic, %	Fail-safe N	Regression intercept P value
Overall bullyingd
End of intervention	32.6 (23.7 to 41.6)	NA	45	46 847	45 744	−0.150 (−0.191 to −0.109)	<.001	<.001	85.3	209	.03
Follow-up	31.5 (15.8 to 47.2)	44.0 (3 to 104)	21	11 020	11 977	−0.171 (−0.243 to −0.099)	<.001	<.001	80.0	16	.09
Bullying perpetration
End of intervention	35.9 (25.3 to 46.6)	NA	35	43 199	42 991	−0.111 (−0.146 to −0.077)	<.001	<.001	78.8	558	.006
Follow-up	33.4 (15.4 to 51.4)	39.2 (3 to 104)	17	7889	7993	−0.175 (−0.276 to −0.073)	.002	<.001	85.9	49	.18
Bullying exposure
End of intervention	34.8 (22.9 to 46.6)	NA	32	37 190	37 001	−0.158 (−0.225 to −0.092)	<.001	<.001	94.1	25	.33
Follow-up	23.5 (9.5 to 37.5)	40.9 (4 to 78)	13	6971	7629	−0.122 (−0.173 to −0.071)	<.001	.060	41.3	12	.20
Cyberbullyinge
End of intervention	33.4 (5.5 to 67.3)	NA	5	3271	2472	−0.135 (−0.201 to −0.069)	<.001	.290	19.7	5	.34
Follow-up	78.0 (NA)	52.0 (NA)	1	NA	NA	NA	NA	NA	NA	NA	NA
Attitudes that discourage bullying
End of intervention	27.7 (19.7 to 35.6)	NA	25	20 537	17 778	0.195 (0.145 to 0.245)	<.001	<.001	78.4	4	.007
Follow-up	34.8 (17.4 to 52.2)	50.1 (2 to 104)	14	5517	4596	0.143 (0.083 to 0.202)	<.001	.011	52.5	2	.06
Attitudes that encourage bullying
End of intervention	27.1 (17.4 to 36.8)	NA	15	15 884	14 037	−0.115 (−0.184 to −0.046)	.04	<.001	85.2	14	.58
Follow-up	19.2 (10.8 to 48.2)	48.6 (4 to 78)	7	3329	3299	−0.123 (−0.197 to −0.048)	.002	.070	48.6	69	.69
Mental health problems
End of intervention	25.7 (11.1 to 40.2)	NA	20	14 543	14 649	−0.205 (−0.277 to −0.133)	<.001	<.001	83.7	10	<.001
Follow-up	20.8 (7.8 to 44.4)	27.3 (2 to 52)	6	1605	1621	−0.202 (−0.347 to −0.056)	.01	.012	65.7	4	.001
School climate
End of intervention	36.5 (13.1 to 59.9)	NA	12	11 417	11 995	0.070 (0.044 to 0.096)	<.001	.700	0	NA	.02
Follow-up	18.8 (8.3 to 44.9)	62.4 (52 to 78)	5	2647	2978	0.135 (0.037 to 0.233)	.009	.006	72.0	1	.92

Abbreviations: FDR, false discovery rate; k, number of samples; NA, not applicable.

^aPositive Cohen d values indicate that a specific variable increases more in the intervention group than in the control group during the assessed period, whereas negative Cohen d values indicate the opposite.

^bQ statistic is a magnitude of heterogeneity, and I² statistic is a measure of the proportion of variance in summary effect size attributable to heterogeneity.

^cFail-safe N by Orwin³⁹ refers to the number of unpublished studies required to move estimates to a nonsignificant threshold. Regression intercept P value refers to the linear regression method by Egger et al.⁴⁰

^dOverall bullying is a pooled measure, including data on bullying perpetration, bullying exposure, and cyberbullying.

^eCyberbullying includes pooled cyberbullying perpetration and cyberbullying exposure data.