. 2020 Mar 10;11:383. doi: 10.3389/fpsyg.2020.00383

TABLE 4.

Univariate test results and Bonferroni’s adjusted post hoc pair-wise comparisons from MANOVA between the three groups of children (SRG, YG, and IG).

	F	Effect sizes	Groups		M_diff	p	d
STM	F(2,147) = 2.42, p = 0.09	0.03	YG	IG	0.40	0.24	0.33
			SRG	IG	0.45	0.15	0.41
			YG	SRG	0.05	1	0.05
WM	F(2,147) = 3.14, p = 0.046	0.04	YG	IG	–0.095	0.04	0.48
			SRG	IG	–0.053	0.50	0.26
			YG	SRG	–0.042	0.89	0.25
Inhibitory control	F(2,147) = 46.16, p ≤ 0.001	0.39	YG	IG	–3.57	≤0.001	1.60
			SRG	IG	–3.55	≤0.001	1.57
			YG	SRG	0.02	1	0.01
Forward counting (1–10)	F(2,147) = 9.73, p ≤ 0.001	0.12	YG	IG	–1.71	≤0.001	0.83
			SRG	IG	–1.11	0.02	0.76
			YG	SRG	0.60	0.43	0.24
Forward counting (11–50)	F(2,147) = 74.93, p ≤ 0.001	0.51	YG	IG	–21.51	≤0.001	2.19
			SRG	IG	–18.19	≤0.001	1.71
			YG	SRG	–3.31	0.29	0.42
Backward counting (10–1)	F(2,147) = 12.97, p ≤ 0.001	0.15	YG	IG	–3.57	≤0.001	0.81
			SRG	IG	–3.70	≤0.001	0.86
			YG	SRG	0.13	1	0.03
Backward counting (50–11)	F(2,147) = 5.39, p = 0.006	0.07	YG	IG	–3.76	0.02	0.49
			SRG	IG	–3.65	0.02	0.47
			YG	SRG	–0.11	1	0.12
Digit-quantity mapping	F(2,147) = 44.81, p ≤ 0.001	0.38	YG	IG	–3.54	≤0.001	1.74
			SRG	IG	–2.70	≤0.001	1.37
			YG	SRG	0.84	0.12	0.42
Digit naming (1–10)	F (2,147) = 64.20, p ≤ 0.001	0.47	YG	IG	–4.65	≤0.001	2.19
			SRG	IG	–3.53	≤0.001	1.59
			YG	SRG	–1.12	0.04	0.50
Digit naming (11–16)	F(2,147) = 42.49, p ≤ 0.001	0.37	YG	IG	–2.42	≤0.001	1.40
			SRG	IG	–2.33	≤0.001	1.34
			YG	SRG	–0.09	1	0.13

The MANOVA results revealed a significant main effect for group factor [Wilks’ Lambda = 0.31, F(20,276) = 11.19, p = ≤ 0.001, $η_{p}^{2}$ = 0.45], since the three groups (Syrian refugees Group-SRG, Yazidi Group-YG, and Italian Group-IG) significantly differed from each other. To compare differences between groups, $η_{p}^{2}$ was used as a measure of effect size. Cohen’s (1988) criteria were used to classify the effect sizes: small effect: $η_{p}^{2}$ = 0.01; medium effect: $η_{p}^{2}$ = 0.06; and large effect: $η_{p}^{2}$ = 0.14. Cohen’s d for post hoc pairwise comparisons were used as a measure of effect size: small effect d = 0.20; medium effect d = 0.50; large effect d = 0.80.