Correlates of self-reported offending in children with a first police contact from distinct socio-demographic and ethnic groups

Table 5 Regression analyses per group

		N	Variables	beta	Sign (p)	R²	Anova F(df), p
Moderate/High SES Dutch¹	Child characteristics⁴	64	ADHD (p) Behavior problem (p)	.302 .261	.019 .042	.231	9.139(2), .000
	Family characteristics	66	Low positive parenting	.310	.023	.092	6.804(2), .011
	Combined model	64	ADHD (p) Behavior problem (p)	.302 .261	.019 .042	.231	9.139(2), .000
Low SES Dutch²	Child characteristics⁵	53	Behavior problem (p) Proactive aggression (c)	.396 .273	.002 .032	.290	10.229(2), .000
	Family characteristics		No variables entered
	Combined model	53	Behavior problem (p) Proactive aggression (c)	.396 .273	.002 .032	.290	10.229(2), .000
Low SES non-Western (excl. diagnoses)³	Child characteristics⁶	94	Sensation seeking Behavior problem (p) Substance use Poor relationship with peers	.236 .224 .254 .213	.012 .023 .005 .025	.356	12.001(4), .000
	Family characteristics	100	Mental health problems parent Not both parents in the home	.298 .250	.002 .009	.160	9.235(2), .000
	Combined model	93	Behavior problem (p) Poor relationship with peers (p) Substance use Parental mental health problems Sensation seeking (c)	.198 .207 .244 .239 .224	.040 .025 .005 .006 .011	.412	12.207(5), .000

Note. (p) parent report
(c) child report
Given the limited sample sizes of both Dutch subsamples, only the strongest correlations were entered in the regression analyses up to a maximum of 5. Collinearity proved not to be a problem. However, to limit overlap between constructs, if both child and parent report on the same SDQ scale were correlated, the strongest one was entered. In both Dutch groups, if both a behavior problem scale and a DBD diagnosis or both a hyperactivity scale and an ADHD diagnosis were correlated, the strongest one was entered.
1. If the analyses were run excluding the psychiatric diagnoses ADHD en DBD, only behavior problems entered the model.
2. Running the regression analyses without the psychiatric diagnoses ADHD and DBD produced the same model.
3. The regression analyses including the psychiatric diagnoses ADHD and DBD produced the same model but in a smaller sample.
4. ADHD, parent report behavioral problems, status offenses, reactive aggression, sensation seeking and proactive aggression were entered in the model.
5. Parent report emotional problems, parent report behavioral problems, proactive aggression, ADHD en status offenses were entered in the model.
6. Sensation seeking, parent report behavioral problems, substance use, parent report poor relationship with peers and child report hyperactivity were entered.

ISSN: 1753-2000