Background: Several studies have investigated the association between male pattern baldness and disease such as prostate cancer and cardiovascular disease. Limitations in the lack of standardized instruments to measure male pattern baldness have resulted in researchers measuring balding patterns in a variety of ways. This paper examines the accuracy and reliability of assessment of balding patterns by both trained observers and men themselves, using the Hamilton-Norwood classification system.Methods: An observational study was carried out in Western Australia with 105 male volunteers aged between 30 and 70 years. Participants completed a short questionnaire and selected a picture that best represented their balding pattern. Two trained data collectors also independently assessed the participant's balding pattern using the same system and the men's self assessment was compared with the trained observer's assessment. In a substudy, observers assessed the balding pattern in a photo of the man aged 35 years while the man independently rated his balding at that age.Results: Observers were very reliable in their assessment of balding pattern (85% exact agreement,kappa = 0.83). Compared to trained observers, men were moderately accurate in their self-assessment of their balding status (48-55% exact agreement,. = 0.39 - 0.46). For the substudy the exact agreement between the men and the observers was 67% and the agreement within balding groups was 87%.Conclusions: We recommend that male balding patterns be assessed by trained personnel using the Hamilton-Norwood classification system. Where the use of trained personnel is not feasible, men's self assessment both currently and retrospectively has been shown to be adequate.