Aim: To establish the association between impaired vision and drivers' decisions to stop driving, voluntarily restrict driving, and motor vehicle accidents. Methods: Driving related questions were included in a population based study that determined the prevalence and incidence of eye disease. Stratified random cluster samples based on census collector districts were selected from the Melbourne Statistical Division. Eligible participants aged 44 years and over were interviewed and underwent a comprehensive ophthalmic examination. The outcomes of interest were the decision to stop driving, limiting driving in specified conditions, and driving accidents. The associations between these outcomes and the legally prescribed visual acuity (<6/12) for a driver's licence were investigated. Results: The mean age of the 2594/3040 (85%) eligible participants was 62.5 (range 44-101). People with visual acuity less than 6/12 were no more likely to have an accident than those with better vision (χ2 = 0.175, p>0.9). Older drivers with impaired vision, more so than younger adults, restrict their driving in visually demanding situations (p<0.05). Of the current drivers, 2.6% have vision less than that required to obtain a driver's licence. The risk of having an accident increased with distance driven (OR 2.57, CL 1.63, 4.04 for distance >31 000 km) but not with age. Conclusion: There was no greater likelihood of self reported driving accidents for drivers with impaired vision than those with good vision. While many older drivers with impaired vision limit their driving in adverse conditions and some drivers with impaired vision stop driving, there are a significant number of current drivers with impaired vision.