The membership diversity of voluntary associations is of central interest in the literature investigating the importance of involvement in voluntary associations for civic life. Due to the limited availability of data concerned with the membership composition of voluntary associations, many researchers have adopted a proxy approach that is based on an aggregation of the characteristics of survey respondents who belong to particular types of associations. However, this proxy approach has not yet been validated to assess whether it actually captures voluntary association membership diversity. We address this gap by comparing the proxy approach with a more direct approach for measuring association diversity by using data from the United States Citizenship, Involvement, Democracy Survey and the Social Capital Community Benchmark Survey. Our analyses reveal that the proxy measures are not correlated with direct measures of voluntary association membership diversity.