Multilevel modeling approaches to the study of LGBT-parent families: Methods for dyadic data analysis

Abstract

While obtaining information from multiple members of a family can enhance researchers— understanding of families, it can also present complications when trying to analyze the data, as most traditional statistical methods assume that data originate from independent sources. An additional problem arises when examining data from partners in same-sex couples, which are often “indistinguishable” as they cannot be distinguished on the basis of some characteristic (e.g., gender) meaningful to the analysis. This chapter introduces approaches to analyzing data from “indistinguishable” partners using multilevel modeling for both cross-sectional and longitudinal analysis. It also discusses ways to examine data from multiple informants—for instance, when both mothers in lesbian-parent families report on their child's well-being. Examples are drawn from the authors— recent projects to illustrate the statistical concepts and difficulties.