. 2021 Jun 5;31(11):1753–1764. doi: 10.1007/s00787-021-01813-7

Table 4.

Gender-specific associations between age at school entry (independent variable) and the dichotomised SDQ hyperactivity/inattention subscale (dependent variable; suspected ADHD with scores > 5) as assessed by logistic regression analysis^a

	Three months after school entry		End of first grade
	Parent reports		Parent reports		Teacher reports
	Odds ratio (95% CI)	p (Interaction)^f	Odds ratio (95% CI)	p (Interaction)^f	Odds Ratio (95% CI)	p (Interaction)^f
Model 1^b
Boys	1.15 (0.47, 2.82)	0.98	0.38 (0.17, 0.84)*	0.34	0.72 (0.40, 1.28)	0.092
Girls	1.13 (0.38, 3.34)	0.98	0.21 (0.08, 0.56)**	0.34	0.30 (0.13, 0.70)**	0.092
Model 2^c
Boys	1.02 (0.41, 2.56)	0.78	0.35 (0.16, 0.78)*	0.27	0.64 (0.35, 1.19)	0.041
Girls	0.82 (0.27, 2.54)	0.78	0.17 (0.06, 0.47)***	0.27	0.21 (0.09, 0.52)***	0.041
Model 3^d
Boys	1.42 (0.51, 3.96)	0.39	0.39 (0.17, 0.93)*	0.092	0.70 (0.37, 1.33)	0.020
Girls	0.70 (0.20, 2.36)	0.39	0.12 (0.04, 0.37)***	0.092	0.19 (0.07, 0.47)***	0.020
Model 4^e
Boys			0.32 (0.13, 0.82)*	0.15	0.67 (0.35, 1.28)	0.027
Girls			0.11 (0.03, 0.36)***	0.15	0.18 (0.07, 0.48)***	0.027

CI confidence interval

*p < 0.05

**p < 0.01

***p < 0.001

^aA marginal logistic model for correlated outcomes (parent reports at T2, T3, teacher reports at T3; assessment) was fitted for each adjustment set, containing assessment and all interactions between assessment and all other variables as fixed effects

^bUnadjusted

^cAdjusted for socio-economic status, migrant background, and family structure

^dAdjusted for socio-economic status, migrant background, family structure, and hyperactivity/inattention subscale at baseline

^eAdjusted for gender, socio-economic status, migrant background, family structure, and hyperactivity/inattention subscale at baseline and 3 months after school entry

^b,c,d,eBased on 1236 parent reports (3 months after school entry), 1127 parent reports (end of 1st grade), and 1393 teacher reports. Missing values in outcomes and adjustment set (c) multiply imputed by MCMC method. Full conditionals specified as logistic models for all dichotomized SDQ hyperactivity/inattention subscale scores, using 100 imputations

^fChi-square test for interaction between gender and age at school entry