All time references are in CEST
The LGBTQI* challenge: How to include sexual and gender minorities in general population, longitudinal and cross-national surveys? 1 |
|
| Session Organisers | Mrs Lisa de Vries (Bielefeld University) Dr Mirjam Fischer (University of Cologne) Professor Stephanie Steinmetz (Universities of Lausanne and Amsterdam) |
| Time | Tuesday 18 July, 11:00 - 12:30 |
| Room | U6-28 |
The ongoing public debate and rising legal acceptance of lesbian, gay, bisexual, trans*, queer, and inter* people increased the visibility of LGBTQI* people in the last decades. In addition, scientific studies from various disciplines increasingly examine the living and working conditions of LGBTQI* people across countries.
Nevertheless, LGBTQI* people are still rarely represented in general population surveys and recent data lacks longitudinal and cross-national perspectives. In many countries, even the amount of LGBTQI* people in the whole population cannot be estimated. Recent developments tried to fill these data gaps by i) integrating questions about sexual orientation and gender identity in general population surveys, ii) identifying same-sex couples in register and census data, or iii) integrating new sampling strategies to reach LGBTQI* people. However, these new approaches pose several methodical challenges for researchers and survey institutes.
This session bundles several methodical issues about including LGBTQI* people in general population, longitudinal and cross-national surveys, and invites researchers as well as employees from survey institutes and census bureaus to discuss recent developments and issues.
Relevant topics include (but are not limited to):
• Identify same-sex, bisexual or transgender people in surveys and register data
• Methods of weighting or alignment to integrate LGBTQI* respondents in general population surveys
• Harmonization of changes in gender and sexuality measurements in panel studies over time
• Sampling strategies and new approaches to reach LGBTQI* people for different types of surveys
• Dealing with small group sizes in statistical analyses
Keywords: Sexual Orientation, Gender Identity, General Population Surveys, Longitudinal Surveys, Cross-National Surveys
Dr Kristen Miller (National Center for Health Statistics) - Presenting Author
In 2022, the National Center for Health Statistics conducted a series of studies to understand question response processes related to the collection of gender as a non-binary construct. The work consisted of both cognitive interviewing studies as well as split-sample experiments involving different question design elements. Additionally, the work consisted of several applied research studies to specifically develop a single non-binary item for a variety of data collection systems, including the U.S. Passport, the State Department’s employee management system, and NCHS’s population-based health surveys. This presentation describes the various studies and discusses more broadly how the concepts of purpose, context, privacy along with the need for inclusion, should inform the ways in which the Federal government asks about gender.
Dr Shelley Feuer (U.S. Census Bureau) - Presenting Author
Mr Matt Virgile (U.S. Census Bureau)
Ms Jessica Holzberg (U.S. Census Bureau)
Dr Renee Ellis (U.S. Census Bureau)
Reliable measurement of sexual orientation is crucial to ensuring that marginalized populations are accurately represented. To capture the nuances of sexual identity, sexual orientation questions in surveys often include an open-ended “something else” response option, after closed options for “lesbian or gay,” “straight,” and “bisexual.” However, diversity in who selects “something else” can pose challenges to analyzing and understanding data on sexual orientation. In one Census Bureau survey, most respondents who selected this category used it to assert they are “normal” or object to the question, leading to their exclusion from analyses. Yet, analysis of a national health survey focusing on sexual minority women, showed that most respondents chose “something else” because their sexuality is fluid; these respondents were also younger with unique social and health indicators. Perhaps one way to reduce measurement error in “something else,” therefore, is by examining differences in open-ended “something else” responses for different populations.
This paper addresses the extent to which the type of “something else” responses vary by sociodemographic characteristics in three large-scale federal surveys. We first discuss the development of a coding scheme to categorize open-ended responses by content and sentiment; then we examine the relationship between respondent characteristics such as age, sex, race, and education and the distribution of categories. For instance, we analyze whether younger respondents are more likely to report an alternate sexual identity (and which terms are most frequently used) and whether older respondents are more likely to write “normal” instead of selecting “straight.” We also examine the profile of respondents whose responses show disapproval for the question. The findings suggest potential strategies for measurement of sexual orientation by determining when and for whom “something else” responses can be back-coded into closed-ended sexual minority options, taken seriously as alternate sexual minority identities, or treated as item nonresponse.
Dr Keming Yang (Department of Sociology, University of Durham) - Presenting Author
Intersectionality has become a widely adopted theoretical perspective and an analytical tool in many disciplinary areas in the social sciences. Empirical investigations on this important issue, however, are still trying to catch up with the theoretical development, especially among quantitative research methods for analysing intersectionality. A particular challenge is the potential problem in which a sample drawn to represent the population in a single dimension may turn out to be unrepresentative of certain intersectional groups. This paper aims to explain why this is an issue and discuss potential strategies that may resolve this issue. So far only a handful of studies seem to have considered this issue and offered possible solutions, but they leave some important issues unresolved, such as recruitment of participants, comparability, sampling errors, and response rates. Clearly more work is needed to enable researchers to make the best use of existing sample survey data or to design new sample surveys that could eventually allow them to make meaningful and rigorous connections between sample survey data and the fundamental principles of intersectionality. As this remains a highly exploratory process, the author would like to make the presentation as interactive as possible and welcome any inputs from the audience.
Ms Lisa de Vries (Bielefeld University) - Presenting Author
Dr Mirjam Fischer (Goethe University Frankfurt)
Mr David Kasprowski (DIW Berlin)
Improved social acceptance and legal recognition of gender minorities in many Western countries as well as ongoing public debate increased the visibility and improved living circumstances of gender minorities in the last years. In light of recent developments, German surveys increasingly integrate non-binary sex/gender items into their questionnaires and try to improve the data situation about gender minorities. This improvement comes along with several possibilities but also challenges for current research. In the last years, we integrated different non-binary sex/gender items into two datasets. First, we integrated a nationwide boost sample of sexual and gender minority households (Sample Q) into one of the largest longitudinal household surveys worldwide, the Socio-Economic Panel (SOEP). Second, we conducted an LGBTQI*-Community online survey (LGBielefeld) in the years 2019 and 2021 with about 7,000 respondents for each year. This survey was conducted parallel to the 2019 SOEP in order to allow for comparisons between the two surveys. Based on our experiences our paper compares different approaches to measuring gender and strategies to identify gender minorities in surveys (e.g., 2-step approach and 3-step approach to identify transgender respondents; items with varying numbers of close-ended answer categories to measure gender; an open-ended question to measure gender). We compare these different strategies in terms of their ability accurately to identify gender minorities, their item non-response rates, their reception by respondents, and their efficiency (e.g., response time, effort in data preparation/coding). Our experience with the different measures and their careful comparison will serve as a guide for survey providers and survey users when making decisions about item design or coding decisions of secondary data.
Professor Stephanie Steinmetz (FORS and University of Lausanne) - Presenting Author
Dr Verena Ortmanns (German Institute for Adult Education - Leibniz Centre for Lifelong Learning)
There is an increasing interest among survey providers and population registers/census bureaus to better account for the reality of gender diversity in their data collections. 39 out of 47 Council of Europe member countries currently have a legal gender recognition procedure in place that provides trans and gender-nonconforming people with some degree of legal gender recognition. Therefore, it is important to discuss this inclusion in data collections. Addressing gender diversity in a survey poses challenges to the survey design and instruments on respondents’ sex and gender identity. It might also affect respondents’ answer behavior on other questions.
Against this background, this paper aims at answering the question what are the consequences of adding a third sex designation (e.g. ‘diverse’) for survey quality (item/unit non response, drop out) as well as response patterns on attitudinal gender and non-gender-related questions? Based on probability-based Swiss sample, which has been collected in the framework of the MOSAiC module 2022 on ‘Family and Changing Gender Roles’ we implemented an innovative experimental design which allows us to shed light on these under researched aspects. The expected findings are not only scientifically relevant, but will also contribute to the ongoing societal debate on this topic in Switzerland on whether and how to better recognize transgender and non-binary people. To our knowledge, this is the first study examining in more detail the advantages and challenges related to integrating gender diversity, especially with regard to the question of respondents’ sex, in a representative Swiss population survey. Accounting for the regional diversity within Switzerland, we might even arrive at some new insights.