It's difficult for scientists to tell whether or not people are giving truthful answers in their studies, but sometimes it's quite easy to determine when they're not.
One study recently looked at virginity pledges in the same population of students over a multiple-year period and found some very interesting results. Half of all students who said they had taken a virginity pledge in the first wave of the study denied that pledge in the second wave one year later. The people who were most likely to deny their initial statement that they'd taken a virginity pledge? Young, African-American men who had started to have sex in the interval between surveys.
How do scientists deal with not knowing whether or not people are telling the truth? Well, that's one of the reasons why results are always estimates! However, scientists can use some techniques to improve the quality of their data. People have, for instance, been found to give seemingly more realistic answers about sensitive information on computer-based surveys then they do in face-to-face interviewers. For example, more people admit to drug use and controversial sexual behaviors when they can just type their responses on an anonymous machine than if they have to admit them to a person. Scientists can take advantage of this fact, as well as other research techniques, to try and get the most accurate information possible. Still, no single study is perfect and that's why it takes years of research, and numerous published reports, before most scientists will consider there to be sufficient evidence for a policy, or procedural, change.
Sources:
Rosenbaum JE. "Reborn a virgin: adolescents' retracting of virginity pledges and sexual histories." Am J Public Health. 2006 Jun;96(6):1098-103.
Turner CF, Ku L, Rogers SM, Lindberg LD, Pleck JH, Sonenstein FL. Adolescent sexual behavior, drug use, and violence: increased reporting with computer survey technology. Science 1998 May 8;280(5365):867-73.
