Table 3.

Logistic regression models and AUC values for ASPD prediction by each informant’s total aggression score or separate direct and indirect aggression scores, adjusted for sex and age

Informant (age of participant)	Total aggression models						Direct + indirect aggression models
	N	Model variable	OR	95% CIs	AUC	95% CIs	N	Model variable	OR	95% CIs	AUC	95% CIs
Parent (12)	1278				0.65	0.59, 0.72	1277				0.66	0.60, 0.72
		Total AGG	1.3*	1.1, 1.6				Direct AGG	1.4*	1.1, 1.8
		Sex	2.2*	1.3, 3.8				Indirect AGG	0.9	0.7, 1.1
								Sex	2.0*	1.2, 3.5
Teacher (12)	1302				0.69	0.62, 0.75	1292				0.70	0.63, 0.76
		Total AGG	1.6*	1.3, 1.9				Direct AGG	1.9*	1.4, 2.6
		Sex	1.8*	1.1, 3.0				Indirect AGG	0.8	0.6, 1.2
								Sex	1.4	0.8, 2.5
Teacher (14)	1013				0.72	0.65, 0.80	994				0.73	0.65, 0.81
		Total AGG	1.7*	1.4, 2.1				Direct AGG	1.7*	1.2, 2.3
		Sex	1.7	0.9, 3.2				Indirect AGG	1.1	0.7, 1.5
								Sex	1.6	0.8, 3.2
Self (14)	1312				0.72	0.66, 0.79	1307				0.72	0.66, 0.79
		Total AGG	1.8*	1.5, 2.2				Direct AGG	1.5*	1.2, 1.8
		Sex	1.4	0.8, 2.4				Indirect AGG	1.4*	1.1, 1.7
								Sex	1.4	0.8, 2.4
Co-Twin (14)	1208				0.69	0.62, 0.76	1201				0.69	0.62, 0.76
		Total AGG	1.6*	1.3, 2.0				Direct AGG	1.7*	1.3, 2.2
		Sex	1.6	0.9, 2.9				Indirect AGG	1.0	0.8, 1.3
								Sex	1.5	0.8, 2.6

AGG aggression, ASPD antisocial personality disorder, AUC area under the (receiver operating characteristic) curve, CI confidence interval, OR odds ratio

*p < 0.05