Measurement, methods, and divergent patterns: Reassessing the effects of same-sex parents

Abstract

Scholars have noted that survey analysis of small subsamples—for example, same-sex parent families—is sensitive to researchers’ analytical decisions, and even small differences in coding can profoundly shape empirical patterns. As an illustration, we reassess the findings of a recent article by Regnerus regarding the implications of being raised by gay and lesbian parents. Taking a close look at the New Family Structures Study (NFSS), we demonstrate the potential for misclassifying a non-negligible number of respondents as having been raised by parents who had a same-sex romantic relationship. We assess the implications of these possible misclassifications, along with other methodological considerations, by reanalyzing the NFSS in seven steps. The reanalysis offers evidence that the empirical patterns showcased in the original Regnerus article are fragile—so fragile that they appear largely a function of these possible misclassifications and other methodological choices. Our replication and reanalysis of Regnerus’s study offer a cautionary illustration of the importance of double checking and critically assessing the implications of measurement and other methodological decisions in our and others’ research.