TABLE 2

Logistic Regression Predicting Antisocial Behavior in Early Adulthood From Television Viewing Between 5 and 15 y Old

Outcome VariableStudy GroupUnadjusted Odds Ratioa (95% Confidence Interval)Multiadjusted Odds Ratiob (95% Confidence Interval)
Any criminal conviction by age 26 yTotal sampleb1.55 (1.26–1.89)**1.27 (1.00–1.61)*
Males1.42 (1.12–1.79)**1.24 (0.93–1.64)
Females1.92 (1.32–2.79)**1.46 (0.93–2.31)
Any violent conviction by age 26 yTotal sampleb1.34 (1.07–1.70)*1.25 (0.95–1.66)
Males1.43 (1.10–1.86)**1.29 (0.93–1.77)
Females1.07 (0.64–1.79)1.10 (0.59–2.06)
Diagnosis of antisocial personality disorder at either age 21 or 26 yTotal sampleb1.70 (1.24–2.34)**1.61 (1.10–2.36)*
Males1.69 (1.21–2.36)**1.62 (1.08–2.43)*
Females
  • Odds ratios indicate the increase in the odds for the outcome associated with a 1-h increase in mean weekday television viewing. Significant at *P < .05, **P < .01. —, Too few cases to analyze (n = 4).

  • a Odds ratios are unadjusted for male and female subgroups but adjusted for sex for the total sample.

  • b Adjusted for sex, childhood SES, childhood IQ, undercontrolled temperament at age 3 y, parent- and teacher-rated antisocial behavior at age 5 y, and parental control at ages 7 and 9 y.