Purpose <p>Participation rates in population-based health surveys have been declining for decades, with nonresponse bias being a serious threat to the validity and generalizability of results. The aim of the study was to examine differences between participants and nonparticipants invited to a psychiatric diagnostic interview survey in terms of sociodemographic and health characteristics, and use of health services for mental health problems, including diagnoses.</p> Methods <p>The study sample for the follow-up psychiatric diagnostic interview was recruited among participants in the fourth survey of the Trøndelag Health Study (HUNT4) in Norway. Information about sociodemographic and health characteristics was gathered from the main survey, while data on health service use was collected by linking records from primary and specialist patient registries with information about participation status.</p> Results <p>Male sex, younger age, being unmarried, having lower educational attainment, and having lower income were associated with higher odds for nonparticipation. Contact with the primary or specialist health services for mental health problems and hospitalization for a mental disorder more than a year before or after invitation date, were associated with lower odds for nonparticipation, especially for diagnoses indicating affective disorders, anxiety disorders, personality disorders, hyperkinetic disorders or milder mental complaints.</p> Conclusion <p>Contrary to most prior studies examining nonresponse bias, the results indicate that people who had been in contact with the health services for mental health problems were more inclined to participate in a survey focused on mental disorders. The direction of nonresponse bias should be carefully considered and adjusted for in population-based studies.</p>