The NHLBI Twin Study is a longitudinal study of cardiovascular disease risk factors in 514 pairs of white, middle aged, male, veteran twins. The initial examination took place between 1969-1973. Ten years later, 81% of the living cohort returned for a second examination. Data collected up to 30 years prior to recruitment for the initial examination were used to characterize participants and nonparticipants; data from the initial examination were used to characterize returnees and nonreturnees to the second examination. Participants had significantly lower diastolic blood pressure and higher socioeconomic status than nonparticipants as measured thirty years earlier. Between the first and second examinations, the mortality of participants was less than 50% of the mortality of nonparticipants. Returnees to the second examination had a better health profile at the initial examination than nonreturnees, with significantly lower levels of cigarette smoking, glucose intolerance, hypertension, and diabetes and higher levels of pulmonary function. However, returnees were more obese than nonreturnees. Thus, this study of cardiovascular disease risk factors in twins appears to be affected by response bias in a way similar to studies of individuals. Additional analyses of biases that may affect the genetic component of the study indicated that factors related to classical twin analyses were relatively unaffected by selection.