There is substantial disagreement among published epidemiological studies regarding environmental risk factors for Parkinson's disease (PD). Differences in the quality of measurement of environmental exposures may contribute to this variation. The current study examined the test-retest repeatability of self-report data on risk factors for PD obtained from a series of 32 PD cases recruited from neurology clinics and 29 healthy sex-, age- and residential suburb-matched controls. Exposure data were collected in face-to-face interviews using a structured questionnaire derived from previous epidemiological studies. High repeatability was demonstrated for 'lifestyle' exposures, such as smoking and coffee/tea consumption (kappas 0.70-1.00). Environmental exposures that involved some action by the person, such as pesticide application and use of solvents and metals, also showed high repeatability (kappas>0.78). Lower repeatability was seen for rural residency and bore water consumption (kappa 0.39-0.74). In general, we found that case and control participants provided similar rates of incongruent and missing responses for categorical and continuous occupational, domestic, lifestyle and medical exposures.