Although hearing loss as a result of aging is thought to be inevitable, in industrialized countries it has been accelerated by excessive exposure to nonoccupational and occupational noise (Phaneuf & [...], The New York State Farm Family Health and Hazard Surveillance was conducted to assess the health status and safety practices among year-round adult farmers and farm residents in New York State and included a telephone interview survey of 1,727 persons from 552 farms. To determine the extent to which self-reported hearing loss is in agreement with audiometry, a subset of 376 participants who completed a hearing loss interview and pure-tone audiometry was analyzed. Thirty-six percent of the participants had self-reported hearing loss, defined as at least some difficulty hearing in one or both ears. The prevalence of audiometric hearing impairment, defined as a threshold average greater than 25 dB hearing level, was 9% for the binaural low-frequency average (500, 1000, and 2000 Hz), 29% for the binaural mid-frequency average (1000, 2000, 3000, and 4000 Hz), and 47% for the binaural high-frequency average (3000, 4000, 6000, and 8000 Hz). Agreement between self-report and audiometry was highest for the binaural mid-frequency average (kappa statistic 55%, sensitivity 77%, and specificity 82%). Self-reported hearing loss was found to be a moderately good measure of hearing impairment. We conclude that a simple questionnaire focusing on hearing difficulty is a useful and valid tool for conducting epidemiologic studies of farmers. Whenever possible, a substudy using audiometry should be conducted. KEY WORDS: hearing loss, audiometry, questionnaire, agreement