While sorting through the NLSY79, I realized women often answer one particular question inconsistently from year to year.
How old were you when you first had sexual intercourse?
A subset of the 79 cohort was asked this question in 1983, 84, and 85. More often than not, they answered inconsistently. 1 year differences could probably be chalked up to natural error. "Was I 16 or 17 that Summer?" What do you say for a 16/20 split though? Clearly, someone is not telling the truth. But why lie to an interviewer?
After skimming the NLSY79 manual, it appears as the interviewee was not always surveyed alone (spouses, children, etc. were present). That's unfortunate, from a scientific perspective. I plan to use the lowest year offered by the interviewee. What do you think?
(The question was not repeated for males once they had offered a year, so comparisons are not possible.)