In preliminary analyses, we examined bivariate associations between exposure to sexual content on television and pregnancy and between pregnancy and each covariate.
We also used logistic regression analyses to test the partial association between baseline exposure to sexual content on television and pregnancy after baseline, with adjustment for total television exposure. Our previous work18,34 indicated that youths who see more sexual content on television are also youths who watch more television overall; failure to account for this association results in sexual content exposure acting in part as a “stand-in” for overall viewing. This partial correlation better captures the “true” relationship between sexual
content and pregnancy (before adjustment for demographic and other individual characteristics)