Abstract
This study examined the agreement between daily and retrospective reports of vaginal sex over a two-month period in a sample of 376 heterosexually active men and women. We also examined whether gender, age, or method of daily data collection (self-administered vs. interviewer administered) were related to agreement between daily and retrospective reports. Both counts and categorical measures of frequency of the behaviors were examined. There were no gender, age, or data collection method effects. When measured as a count, participants reported more instances of vaginal intercourse in the retrospective reports than on the daily reports. In contrast, comparison of retrospective categorical measures of frequency to daily reports showed considerable variability. Possible reasons for the over-reporting of counts of vaginal sex in retrospective reports are explored.
This work was supported by grants from the National Institute on Alcohol Abuse and Alcoholism (R01AA013688, R01AA000183, and K05AA015050).
Notes
1 This adjustment assumes that the probability of sexual behavior is equal on missing and non-missing days. If not, the difference between diary and retrospective measures is incorrect—that is, if someone was more likely to have sex on missing days than on non-missing days, the daily measure is biased downward, while the retrospective measure is not; the discrepancy between diary and retrospective measures would be an artifact of missing days.