Background: Population-based estimates of biological family size can be useful for planning genetic studies, assessing how distributions of relatives affect disease associations with family history and estimating prevalence of potential family support. Methods: Mean family size per person is estimated from a population-based telephone survey (n = 1,019). Results: After multivariate adjustment for demographic variables, older and non-White respondents reported greater mean numbers of total, first- and second-degree relatives. Females reported more total and first-degree relatives, while less educated respondents reported more second-degree relatives. Conclusions: Demographic differences in family size have implications for genetic research. Therefore, periodic collection of family structure data in representative populations would be useful.