ESRA logo

ESRA 2023 Glance Program

All time references are in CEST

Developments in survey methods and analysis about LGBTI+ populations 2

Session Organiser Dr Angelo Moretti (Utrecht University)
TimeWednesday 19 July, 14:00 - 15:00
Room U6-23

Empirical evidence is crucial in shaping and implementing policies focusing on lesbian, gay, bisexual, transgender, transexual, intersexual and any other individuals whose sexual and/or gender identity differs from the cis-heterosexual (LGBTI+). In recent years, much progress has been made regarding the legislation focusing on the protection and support of LGBTI+ communities. However, obstacles are still present in practice when it comes to data collection and analysis to provide reliable outputs to be used by policy makers. Particularly, there remains gaps in data collection initiatives aiming to investigate a wide range of issues affecting LGBTI+ populations. This lack of data affects policies, and this has an important impact on the lives and well-being of LGBTI+ people. These populations can be seen as ‘hard-to-reach populations’, meaning that they are difficult for survey researchers to access; hence, relevant sampling designs should be considered and formulated carefully. Furthermore, there are important methodological issues arising from when collected data are analysed. Specifically, confidentiality and privacy issues are crucial here. This can lead onto measurement error issues in the reporting stage, e.g., discriminations in various contexts, and victimisations. Thus, these problems must be taken into account when developing and applying statistical models. In addition, due to the nature of this phenomena object of study, distributions might suffer from a large number of zeros so that the data do not readily fit standard distributions, e.g., variables measuring victimisation of LGBTI+ people. We welcome substantive and methodological papers that address issues related to survey methods and analysis of LGBTI+ populations. Methodological papers can be related to sampling designs, data integration developments considering new forms of data too, statistical modelling approaches that seek to produce robust analysis.

Keywords: lgbti+, lgbt, elusive populations, hard-to-reach, measurement


A propensity weighting approach for nonprobability samples

Dr Ernest Albert (University of Vienna)
Dr Karel Fric ( - Presenting Author

When the EU Agency for Fundamental Rights (FRA) carried out their second LGBTI survey in 28 EU Member states, North Macedonia and Serbia, almost 140'000 respondents replied to the survey and provided information on their experiences of discrimination, unsafe environments, hate motivated violence, harassment, and on other crucial topics for a better realisation of the fundamental rights of LGBTI people in Europe. Despite the success in absolute numbers of recruited participants, a challenge remained for the data analysis: accounting in the best possible way for the self-selection bias, on several dimensions, expectable in a nonprobability sample of this type.

One of the measures applied in 2019 has been to adapt propensity weighting to a nonprobability sampling environment by (a) using self-reported participation in an earlier, similar survey as a proxy for the propensity to participate in the present survey; and by (b) empirically linking this propensity to key variables associated with the recruitment process. Among these key variables, individually reported recruitment channels have been tested, and – with more success – the affiliation with organisations involved in the recruitment. Details of the approach and empirical evidence that encouraged its application in the FRA LGBTI II survey wave will be presented in the talk. Different measurement error scenarios will be discussed regarding their effect on such propensity weights using the data from 2019. Limitations of the approach will be shown, particularly when a survey does not satisfy certain conditions. On the other hand, the approach's potential will be considered also for other hard to reach populations covered by institutions engaged in the research on minorities.

„Free Expression of One’s Own Personality?“ The Role of Gender Identity and Gender Expression in Poverty Risks.

Mr David Kasprowski (DIW (SOEP), BGSS) - Presenting Author

Mostly, gender is treated as a control variable or an important interaction without any theorization. In addition to strong economic differences based on gender, such as the persistent gender pay gap to the disadvantage of women, it is still not specified which aspects of gender may contribute to disadvantages. However, recent research suggests that self-identified gender non-conforming individuals experience distinctive economic penalties. Based on a German online-survey, this contribution compares the multidimensional assessment of gender with a large sample of 6,956 LGBTI* individuals, including 1,131 transgender, gender non-conforming and non-binary individuals. Beyond the question of the poverty risks of trans* and non-binary individuals compared to cisgender women and men, I particularly address the role of gender expression and perceived gender on social stratification. However, what happens when gender identity and expression do not match and people are ascribed a different gender identity than they have chosen for themselves? Does gender (non) conforming presentation possibly have stronger negative consequences than the gender identity itself? To what extent do common gender theories need to be adapted accordingly or which aspects need to be examined more clearly if we want to understand gendered differences and ultimately reduce discrimination? First results show that especially self-identified non-binary and gender-non-conforming people have to face the highest poverty risk factors like insecure or inadequate housing conditions. The results will be discussed with regard to the theoretical deduction of gender in quantitative research and which possibly false conclusions are drawn when gender is applied as a fixed entity of two genders for analyses without a critical evaluation.

Sports Participation and Sex-Related Homophily in Male Teenagers: Longitudinal Evidence on a Potential Peer Effect from Participating in Male-Dominated Sports

Mr Simon Lütkewitte (Bielefeld University) - Presenting Author

Until today several scholars argue that sports is a social sphere in which hegemonic masculinities continue to be reproduced. Often, this existence of hegemonic masculinities within the sphere of sports is explained by identity processes, meaning participating in sports is argued to foster the development of male-identities. In the literature on sport participation and sex-related attitudes and beliefs, so far no empirical analysis has been published that focuses on potential effects from participating in sports on such identity processes using longitudinal data. However, the data set “Friendship and Identity in School” (FIS) by the German Center for Integration and Migration Research (DeZIM) now offers the possibility to examine the potential effects from participating in different kind of sports on sex-related homophily - describing the tendency to be more likely to bond with people of the same sex - within young teenagers over time. These young teenagers are 11-17 years old and come from nine secondary schools in North Rhine-Westphalia (Germany) and they took part in up to six waves within the years 2013-2017. Besides questions on sport participation and the item on sex-related homophily, the data set also offers the possibility to study the friendship networks of these teenagers over time. So, for the group of male teenagers this work aims to answer the following two research questions: a) whether the participation in male-dominated sports fosters sex-related homophily and b) whether it leads to a higher share of having same-sex-friends. To answer these two research questions fixed-effects models are used. Concerning the first questions, first results revealed that participating in male-dominated sport groups (football, basketball, or martial arts) indeed leads to significantly higher levels of sex-related homophily. The results regarding to the second research question will follow soon.