Counting Same-Sex Unmarried Partners in the 2000 US Census Data: How Many Are Gay and Lesbian?

Dan Black, Syracuse University
Gary J. Gates, Urban Institute
Seth G. Sanders, University of Maryland
Lowell Taylor, Carnegie Mellon University

Large increases in the reporting of same-sex unmarried partners in the Decennial Census 2000 provides a significantly larger sample of same-sex unmarried partners than was available in 1990 and bodes well for research on the demographic and economic aspects of the gay and lesbian population. However, a change in Census coding procedures between 1990 and 2000 creates a problem whereby a heterosexual married couple household that miscodes the sex of one of the partners will now be counted as a same-sex unmarried partner household. Even a small fraction of sex miscodes among married couples creates a sizable measurement error issue among same-sex unmarried partners. We use the Census 2000 Supplementary Survey Public Use Microdata Sample to test a set of models for assessing the extent of the measurement error and calculating reliable estimates of demographic characteristics of gay and lesbian couples.