Visualizing Multiple Sexual Identity Responses Using a Network Approach

Abstract

Measurement strategies for gauging sexual identity within the United States’s population-based data infrastructure are evolving. Despite notable advancements, most measurement techniques still do not accurately reflect the nuanced ways in which sexually diverse adults conceptualize their identities. One measurement limitation involves asking participants to choose a single “best” identity. In this visualization, drawing on data from the National Couples’ Health and Time Study, the authors illustrate the complexity of sexual identity responses by constructing a network that aggregates the patterns of multiple identities chosen by respondents. The findings underscore that relying on a single-item response risks undercounting the range of sexually diverse identities. As national surveys progress in incorporating sexual identity measures, a “select all that apply” option is crucial if the identities of sexually diverse adults are to be accurately represented.

Researchers have refined measures of sexual identity (Bates, Chin, and Becker 2022), facilitating critical research on sexually diverse populations in the United States. Despite considerable progress, limitations remain that do not align with how many sexually diverse adults conceptualize their identity. For instance, the forced-choice selection of a single “best” response option assumes that all respondents identify with a singular identity listed, failing to capture the complexities of sexual identity (Suen et al. 2020). The National Couples’ Health and Time Study (NCHAT) is a population-representative sample of partnered adults between the ages of 20 and 60 years old with an oversample of LGBT adults (Kamp Dush et al. 2023). In the NCHAT, there are over 90 combinations of responses to sexual identity, and 30 percent of respondents (n = 396) that report they are not exclusively heterosexual/straight selected more than one identity (Julian, Manning, and Kamp Dush 2024; Kamp Dush and Manning 2022).

This visualization (Figure 1) illustrates the complexities of sexual identity responses by creating a network that aggregates the patterns of selected identities for individuals who report more than one sexual identity. Networks composed of connections based on shared membership are referred to as “affiliation networks” (Breiger 1974). In our affiliation network, each node represents a sexual identity response, including heterosexual/straight, gay/lesbian, same gender loving, queer, pansexual, omnisexual, asexual, don’t know, questioning, demisexual, and another term. Each edge (or line) connecting two nodes represents respondent(s) selecting those two response categories (respondents may select more than two categories). To focus on the strongest selection patterns, we only include edges where there are at least five respondents who selected the two categories. Edges are weighted to account for the differences in group size, with the thickest edges being the most selected combinations. Nodes (the circles) are sized by the number of responses for each sexual identity, with the largest nodes being the most selected in the network. See the supplemental material for details on the method of analysis.

Figure 1.

Affiliation network of multiple sexual identity responses.

Results

As shown by the size of the nodes in Figure 1, the three most selected identities in the network are bisexual, gay/lesbian, and queer. As shown through the thickest edges in the affiliation network, the most common combinations of two identities involved queer, which included combinations of gay/lesbian and queer and bisexual and queer. Strong selection patterns are also visible between pansexual and bisexual, gay/lesbian and same gender loving, bisexual and gay/lesbian, and heterosexual/straight and bisexual. The most connected identity is bisexual, indicating it was connected alongside of most other identities in the network.

Implications

Our findings show that relying on a single-item response risks undercounting the range of sexually diverse identities. About 3 out of 10 individuals in our sample with a sexually diverse identity selected more than one identity. In our affiliation network, the edges connecting the sexual identity nodes constitute links that would otherwise be missed in measurement strategies that rely on a single “best” response. Furthermore, identities that are often combined with others, such as queer, could be rendered invisible. As national surveys continue to integrate sexual identity measures, a “select all that apply” option is an important inclusion that does not add to the number of survey questions but has the benefit of representing the identities of sexually diverse adults most accurately.

Supplemental Material

sj-docx-1-srd-10.1177_23780231241279129 – Supplemental material for Visualizing Multiple Sexual Identity Responses Using a Network Approach

Supplemental material, sj-docx-1-srd-10.1177_23780231241279129 for Visualizing Multiple Sexual Identity Responses Using a Network Approach by Christopher A. Julian, Lauren Newmyer, Wendy D. Manning and Claire M. Kamp Dush in Socius