Skip to main content

Table 3 Logistic regression model for school attendance status by different groups of substance users

From: A prospective study of methamphetamine use as a predictor of high school non-attendance in Cape Town, South Africa

Dropout/absent

n

Odds ratio

95% CI

(Reference category = Never used meth, never used cannabis and not current smoker)

974

  

   Current smoking, but never used meth or cannabis

150

2.16**

1.44 - 3.25

   Current smoking + life-time cannabis use, but never used meth

125

1.94*

1.09 - 3.47

   Current smoking + life-time cannabis use + life-time meth use

47

2.58*

1.24 - 5.36

Age > than norm for grade

(Ref. Age normal)

1265

2.16**

1.60 - 2.93

Ever repeated a year at school

(Ref. Never repeated a year)

1265

1.32

0.98 - 1.79

  1. * Significant at p < 0.05 level
  2. ** Significant at p < 0.01 level
  3. p-values were from Wald t-tests, df = 14
  4. Note: Some cases excluded from model owing to missing values