BACKGROUND: Recently reported prevalences of myopia in primary school children vary greatly in different regions of the world. This study aimed to estimate the prevalence of refractive errors in an unselected urban population of young primary school children in eastern Sydney, Australia, between 1998 and 2004, for comparison with our previously published data gathered using the same protocols and other Australian studies over the last 30 years. METHODS: Right eye refractive data from non-cycloplegic retinoscopy was analysed for 1,936 children aged 4 to 12 years who underwent a full eye examination whilst on a vision science excursion to the Vision Education Centre Clinic at the University of New South Wales. Myopia was defined as spherical equivalents equal to or less than -0.50 D, and hyperopia as spherical equivalents greater than +0.50 D. RESULTS: The mean spherical equivalent decreased significantly (p < 0.0001) with age from +0.73 +/- 0.1D (SE) at age 4 to +0.21 +/- 0.11D at age 12 years. The proportion of children across all ages with myopia of -0.50D or more was 8.4%, ranging from 2.3% of 4 year olds to 14.7% of 12 year olds. Hyperopia greater than +0.50D was present in 38.4%. A 3-way ANOVA for cohort, age and gender of both the current and our previous data showed a significant main effect for age (p < 0.0001) but not for cohort (p = 0.134) or gender (p = 0.61). CONCLUSIONS: Comparison of our new data with our early 1990s data and that from studies of over 8,000 Australian non-clinical rural and urban children in the 1970's and 1980's provided no evidence for the rapidly increasing prevalence of myopia described elsewhere in the world. In fact, the prevalence of myopia in Australian children continues to be significantly lower than that reported in Asia and North America despite changing demographics. This raises the issue of whether these results are a reflection of Australia's stable educational system and lifestyle over the last 30 years.