We compared self- and spouse reports of snoring and other symptoms of sleep apnea syndrome ascertained from married couples in a community-based survey. Agreement between the two types of report varied between 70-98%, but was modest (kappa = -0.01-0.52) when adjusted for chance. For men, spouse reports yielded higher prevalence rates for snoring and for four other symptoms. For women, estimates of symptom prevalence were consistently lower by spouse report than by self-report. In multivariate analyses, the effect on snoring of gender and obesity increased and of age decreased when spouse reports were compared to self-reports. Snoring, according to spouse reports, was a significant risk factor for ischemic heart disease, but snoring according to self-reports showed a smaller effect and was not statistically significant. Snoring was not associated with hypertension when defined by either self- or spouse report. These observations suggest that questionnaire data of snoring and other symptoms of sleep apnea syndrome may be misclassified in part, and that such misclassification can affect estimates of prevalence and effects.