Revisiting the Generations and Gender Survey: Has Time Improved the Reliability for Studying Same-Sex Couples?

Fernanda Fortes de Lena , Centre d' Estudis Demogràfics
Maaike van der Vleuten, Netherlands Interdisciplinary demographic institute
Lin Rouvroye, Netherlands Interdisciplinary Demographic Institute

Demographic studies on LGBTQ+ populations have grown, yet representative data on queer families remain scarce. The Generations and Gender Survey (GGS) provides valuable cross-national data on family and partnership dynamics but has faced criticism for miscoding and small sample sizes, limiting its reliability for studying same-sex couples. This study reassesses these issues and examines whether the second round of the GGS (Round II) offers improved reliability.Using data from twelve countries in GGS Round II Wave 1 and Wave 2 data from Norway and Estonia, we replicate earlier validation approaches and extend them by considering sexual orientation and the type of child relationship (biological, adopted, or stepchild). We find that countries with legal recognition of same-sex marriage report fewer “false” same-sex couples, suggesting that legal and social visibility enhance data quality. When years of registered partnerships are incorporated into validation, the number of apparent false cases declines markedly—and in some countries with long-standing legal recognition, nearly disappears. This pattern indicates that respondents often view registered partnerships as equivalent to marriage, reflecting both evolving legal frameworks and shifting social understandings of what it means to be “married. While no single method eliminates miscoding, combining legal context, family composition, and sexual orientation data strengthens reliability. Revising the GGS question sequence to ask about registered partnership before marriage—or clarifying that marriage excludes partnerships—would improve comparability and enhance the GGS’s capacity to study diverse family forms.

See extended abstract

 Presented in Session 112. Survey Mode Effects and Measurement Challenges in Demographic Research