Table 4.

Associations of subscales of the Early Child Behavior Questionnaire and receptive language scores at 40 months: regression coefficients in z-score (change in SD), 95% confidence intervals and p-values

	Model 1		Model 2		Model 3
	Coefficient [95%CI]	p	Coefficient [95%CI]	p	Coefficient [95%CI]	p
Motor activation	−0.174 [− 0.254, − 0.095]	< .001	− 0.245 [− 0.339, − 0.150]	< .001	−0.225 [− 0.317, − 0.133]	< .001
Perceptual sensitivity	0.132 [0.069, 0.195]	< .001	0.169 [0.092, 0.245]	< .001	0.150 [0.076, 0.225]	< .001
Attentional shifting	0.178 [0.095, 0.262]	< .001	0.071 [−0.026, 0.168]	.15	0.063 [−0.030, 0.156]	.18
Inhibitory control	0.145 [0.072, 0.218]	< .001	0.082 [−0.001, 0.165]	.05	0.066 [−0.015, 0.146]	.11

Note. Model 1 = Univariate; Model 2 = Adjusted for other subscales of the Early Child Behavior Questionnaire Subscales (ECBQ); Model 3 = Model 2 with further adjustment for child sex, birth weight, gestational age at birth, birth order, age of the mother, years of maternal education, annual household income, maternal history of mood/anxiety disorders. CI Confidence intervals. Bold types represent p < .0013. The regression coefficients shown in Table 4 indicate the amount of predicted change in the z-score of receptive language skills per one unit of change in each of the ECBQ subscales. Again, this implies that a child scoring 7 (maximum) in motor activation, for example, was predicted to score − 0.225 × 6 (7 minus 1 point in the subscale) = − 1.350 points, corresponding to − 1.350 SD higher lower in the receptive language skills score at 40 months compared with a child scoring 1 (minimum)