Participation bias in exercise studies is poorly understood among older adults. This study was aimed at looking into whether older persons who volunteer to participate in an exercise study differ from nonvolunteers. Methods: A self-reported questionnaire on physical activity and general health was mailed out to 1000 persons, aged 60 or over, who were covered by the medical insurance of the French National Education System. Among them, 535 answered it and sent it back. Two hundred and thirty-three persons (age 69.7 +/- 7.6, 65.7% women) said they would volunteer to participate in an exercise study and 270 (age 71.7 +/- 8.8, 62.2% women) did not. Results: Volunteers were younger and more educated than nonvolunteers, but they did not differ in sex. They had less physical function decline and higher volumes of physical activity than nonvolunteers. Compared with volunteers, nonvolunteers had a worse self-reported health and suffered more frequently from chronic pain. Multiple logistic regressions showed that good self-reported health, absence of chronic pain, and lower levels of physical function decline were associated with volunteering to participate in an exercise study. Conclusions: Volunteers were fitter and healthier than nonvolunteers. Therefore, caution must be taken when generalizing the results of exercise intervention studies.
