Skip to main content

Table 3 Logistic Regression Analysis Predicting the Likelihood of Schizophreniaa

From: Combined pattern of childhood psycho-behavioral characteristics in patients with schizophrenia: a retrospective study in Japan

CBCL syndrome subscale

B

SE

p

OR

95% CI for OR

Lower

Upper

Withdrawn

.119

0.038

0.002**

1.127

1.045

1.215

Somatic complaints

−.066

0.040

0.095

.936

.866

1.012

Anxious/Depressed

.031

0.044

0.479

1.032

.946

1.125

Social problems

.035

0.055

0.521

1.036

.931

1.152

Thought problems

.279

0.084

0.001**

1.322

1.122

1.558

Attention problem

.129

0.069

0.061

1.137

.994

1.301

Delinquent behavior

−.011

0.065

0.865

.989

.871

1.123

Aggressive behavior

−.428

0.141

0.002**

.652

.494

.860

  1. ** p < 0.01
  2. aLogistic regression model statistics: Cox-Snell R2 = 0.286, Nagelkerke R2 = 0.439. The goodness-of-fit test of Hosmer and Lemeshow: χ2 = 2.820, df = 7, p = 0.901. Discrimination accuracy = 85.5%. CBCL syndrome scales were all entered into the model as independent variables