Or maybe those teenagers with poor working memories have more unprotected sex?
If the hypothesis works out, how do they know the responses to the survey are accurate?