Table 4

Meta-analysis of variance of moderator variables

DomainAssociationModeratorI2p ValueKESVariance95% CIFSNr
Rating at follow-upAny RP–general BP45.50.02
Expert21.090.100.46 to 1.720−1.00***
Mother140.340.010.21 to 0.49156−0.28
RP assessment methodAny RP–externalising53.410.00
Questionnaire30.220.000.12 to 0.3212−0.76
Interview40.890.050.45 to 1.3215 0.74
Sample characteristicsAny RP–general BP44.020.05
Clinically referred90.610.020.31 to 0.9170−0.23
Community based130.300.000.18 to 0.42143−0.28
  • *** p<0.001.

  • BP, behavioural problems; ES, weighted mean effect size; FSN, Rosenthal's fail-safe N; I2, ratio of true heterogeneity to total variation in observed effects; K, number of studies; r, correlation of sample sizes to effect size (negative correlations indicate that effect size may be overestimated29); RP, regulatory problems.