Prevalence estimates often use U.S. Census Bureau estimates of the population as denominator data. Postcensal estimates are population estimates produced following a decennial census. Intercensal estimates are surrounded by 2 census years and supersede postcensal estimates. In this report we describe prevalence estimates in Atlanta for mental retardation, cerebral palsy, and hearing and vision loss for 8 year olds from 1991-1994 and 1996. We used calculations of postcensal and intercensal population estimates. Intercensal population data were consistently higher than postcensal data, and prevalence estimates for developmental disabilities were lower using intercensal population data. This discrepancy varied by race and ethnicity. Comparison of population estimates, particularly at state and local levels, should be considered to assess meaningful differences in published prevalence estimates using intercensal data.