Table 5.

Results from Analyses of Publication Bias.

Variable	Total number of effect sizes (k)	Published effect size (Hedges' g)	Unpublished effect size (Hedges' g)	Observed effect size (Hedges' g)	Imputed effect size		Eggers test of asymmetry (p-value)	Classic fail-Safe N	Classic-fail-safe N critical value (computed using 5k + 10)
					Duval and tweedie's trim and fill methoda (Hedges' g)	PET-PEESE(Hedges' g)
Incongruent Trials
Accuracy	118	0.273**	0.210*	0.261**	0.261**	−0.330*	<0.001**	2,939	600
Reaction Time	180	0.280**	−0.004	0.239**	0.239**	−0.442*	<0.001**	4,504	910
Congruent Trials
Accuracy	109	0.117	0.093	0.108b	0.108	−0.002	0.865	–b	–b
Reaction Time	181	0.202**	−0.105	0.150*	0.150*	−0.334*	<0.001**	1,688	915
Interference Effect	116	0.335**	0.093	0.268**	0.268**	−0.328*	<0.001**	2,921	590

^*p < 0.05;

^**p < 0.001;

^aMissing studies were assumed to have negative effect sizes (i.e., monolingual evidencing superior performance than bilinguals).

^bClassic Fail Safe N cannot be computed for these tasks since observed average effect size was found to be non-significant.