Skip to main content

Table 6 Risk factors for conviction of offences: Odds ratios and binary logistic regression (forward LR) with 95% CI

From: Child and adolescent psychiatric patients and later criminality

Independent variable

N

Dependent variable

P value

Odds ratio

  

Convicted of offences

No conviction of offences

  

Sex

1400

    

Male

667

367

300

  

Female

733

164

569

 

1.8 < 2.1 < 2.4

Binary logistic regression

1161

  

< 0.001

0.2 < 0.3 < 0.4

Family

1379

    

Split family

724

336

388

< 0.001

1.4 < 1.6 < 1.8

Complete family

655

184

471

  

Binary logistic regression

   

< 0.001

0.4 < 0.5 < 0.6

Problems at school

1170

    

Problems

544

271

273

< 0.001

1.5 < 1.8 < 2.0

No problems

626

177

449

  

Binary logistic regression

   

< 0.001

1.4 < 1.9 < 2.5

Reason for admission to the CAP unit

1389

    

Behavioural disorder

291

179

112

< 0.001

2.7 < 3.5 < 4.5

Other cause for admission

1098

346

752

  

Binary logistic regression

   

< 0.001

2.1 < 3.0 < 4.2

Relationship problems

258

117

141

0.006

1.1 < 1.5 < 1.9

Other cause for admission

1131

408

723

  

Binary logistic regression

   

< 0.001

1.5 < 2.1 < 3.0

  1. P-value and odds ratio from chi-square tests, valid for each variable separately, is shown in plain text. The binary logistic regression odds ratios and p-values are shown in Italic in connection to each variable. See also statistical methods.