Queering the Census: Demographic considerations of adding (and changing) questions on gender and sexuality

Queering the Census: Demographic considerations of adding (and changing) questions on gender and sexuality

“I look at the current precedents and considerations for capturing the LGBTQ+ population in substantively meaningful ways for qualitative and quantitative researchers…”

Christina Pao

Princeton University 

 

Bio

 

Abstract

This paper seeks to answer the following question: Using survey methods, how can we best capture “hidden” populations, and in particular, sexual and gender minorities? Using censuses and national population-based surveys from the UK, US, New Zealand, and Australia, I look at the current precedents and considerations for capturing the LGBTQ+ population in substantively meaningful ways for qualitative and quantitative researchers.
This paper descriptively traces the development and the decision making around gender and sexuality questions on nationally representative surveys and censuses, compares processes, and summarizes some best practices. Some of the initial concerns and challenges at the outset of adding these sociodemographic questions include, from an ethical side: i) privacy concerns and undercounting, ii) the weaponization and politicization of data on marginalized populations, and iii) respondent burden; and, from a methodological side: i) debates on whether queerness can be quantified, ii) survey specific concerns (such as question ordering, response options, question type—like multiple choice or free response, etc…), iii) census specific concerns (such as whether a respondent of a certain age is allowed to answer a sexuality question; whether the head of household responds on behalf of all household members; data protection and transfer between local and national governments, etc…), and iv) comparability/consistency across longitudinal data sets (e.g., whether this meets the needs of data users, particularly from the empirical social sciences, who might be reliant on comparability with past “sex” questions). I will compare and aggregate current guidance and best practice among these four countries.

 

Keywords

census, gender, sexuality, demographics, comparative surveys, population

 

Introduction

Studies show an increase in the number of people who identify with non-binary and non-heterosexual gender and sexual identities, particularly amongst younger cohorts (UK Government Equalities Office 2018; Yurcaba 2022). These population shifts raise important questions for social scientists interested in measuring population inequalities, especially along sex, gender, and sexuality. Social scientists must reconcile concerns of proper comparable measurement against the need to measure social change. This tension has resulted in a growing interest in inclusive but methodologically rigorous measures of gender and sexuality—an interest that has been reflected in changes to many censuses and large-scale social surveys worldwide. 

In light of many countries debating the addition, expansion, and/or implementation of gender and sexuality census questions, this paper seeks to answer the following questions: Through the census, how can we ethically and meaningfully capture “hidden” populations, and, in particular, sexual and gender minorities? What can we learn from the current generative moment, spurred by discourse across and within countries? Using censuses from the US, UK, New Zealand, and Australia, I evaluate some benefits and challenges of adding and/or changing census questions.

Theoretical Motivation

In both the social sciences and colloquial language, discourses around inequality have often centered issues relating to gender and sexuality, whether larger conceptual frameworks or specific areas of discrimination. Studies of inequalities on the basis of gender and sexuality have revealed many robust findings of concern across a variety of disciplines and sectors, and some data has been collected long enough to reveal concerning trends about the potential stagnation that has occurred in attempting to reach equality across these lines of identity (England 2010; Scarborough, Sin, and Risman 2019; Moskos 2020; Loerch 2020; Takács 2015). 

Nonetheless, even with their discursive prevalence, there is often conflation between terms and/or their underlying principles. Both in the social sciences and colloquial language, measures of gender and sexuality are often operationalized inconsistently across data sets, if at all. Because of the issues that arise from entangling sex, gender, and/or sexuality, many governments around the world have begun to reevaluate the questions for these variables on censuses (Hines 2020; Fugard 2020; Harris 2021). I describe some of these changes (or lack thereof) in the following four case studies, the United States, the United Kingdom, Australia, and New Zealand. 

National Census/Survey Background

United States: No present gender/sexuality questions

Of the four countries reviewed in this paper, the US is the only one that maintains a binary sex question without a gender or sexuality question (US Census Bureau n.d.); nonetheless, other large-scale surveys in the US, such as the General Social Survey (GSS), have fielded national surveys with gender and sexuality questions. The lack of change on the US census has been subject to scrutiny from activists (National LGBTQ Task Force 2017).

United Kingdom: Launched gender and sexuality questions

On its most recent census, the UK fielded new gender and sexuality questions with question and response options like those on the US GSS—namely, a two-step gender question and a sexuality question that provided explicit options of straight/heterosexual, gay/lesbian/homosexual, and bisexual (see Table 2 below). Online, the ONS provides its initial 2019 guidance on the sex, gender identity and sexual orientation questions (ONS 2021b)—which was later updated in 2021 due to the High Court ruling in favour of Fair Play for Women (ONS 2021a).

Australia: Expanded sex question, with no gender or sexuality questions

The Australian Bureau of Statistics (ABS), for its 2021 census, expanded the sex question to include an option for a “non-binary sex” (ABS 2021). This was a controversial choice (Karp 2019), and many were upset that the ABS would randomize all “non-binary sex” respondents into “male” and “female” categories for the public release data (Gooley 2021). The 2016 Australian census attempted to capture gender diversity by allowing respondents to access a special online form with an ‘Other’ response option to the Sex question” (ABS 2018). However, because of difficulties with data collection (and likely heightened respondent burden), the 2016 census only counted 1,260 sex and/or gender diverse people in Australia.

New Zealand: Intended gender and sexuality questions, precise wording unknown

New Zealand, after initial testing in 2014, rejected the inclusion of gender and sexual orientation questions since none of the proposed questions had been deemed appropriate for capturing marginalized gender and sexual groups (Strongman 2018); since then, Stats NZ has publicly released new statistical standards for gender, sex, and variations of sex characteristics; added gender and sexuality questions to many national surveys; and committed to adding these updated gender and sexual identity questions to the 2023 census.

Comparative Framework

There have been two main ways that these censuses studied have been attempting to accommodate questions for gender and sexuality: one dimension is by changing questions and another is by adding questions. Some censuses have done both, neither, or one (but not the other). I have displayed the census alterations in a 2×2 matrix below (Table 1). Though these data are not yet released publicly for these most recent censuses, this array of testing across countries provides an interesting source of comparison for the considerations within and between countries in attempting to measure gender and sexuality at a population level. 

Table 1: Comparative Framework Between Australia, New Zealand, the United Kingdom and the United States

		Questions Changed
		Yes	No
Questions Added	Yes	New Zealand (intended)	United Kingdom*
	No	Australia	United States

*Note: The Scottish Census, administered by the National Records of Scotland (NRS), explicitly used different questions than the census administered in England the Wales by the ONS.

Country	Year Implemented	Sex Question	Gender Question	Sexuality Question
Australia	2021	Is this person… (Male/female/non-binary sex)	N/A	N/A
New Zealand	2023	TBD (Might be changed)	TBD (Will be included)	TBD (Will be included)
United Kingdom (ONS)	2021	What is your sex? Note: A questions about your gender will follow if you are aged 16 or over (Female/Male)	Is your gender the same as the sex you were registered at birth? Note: This question is voluntary  (Yes; No, write in gender)	Which of the following best describes your sexual orientation? Note: This question is voluntary (Straight/Heterosexual; Gay or Lesbian; Bisexual; Other sexual orientation, write in)
United States	N/A	What is [Person 1]’s sex? Mark ONE box.  (Male/Female)	N/A	N/A

Table 2: Sex, Gender and Sexuality Census Questions for Australia, New Zealand, the United Kingdom, and the United States 

 (Updated May 2022)

Concerns, Challenges, and Best Practices

At present, these case studies find themselves amid dialogue between data users, policymakers, and LGBTQ+ organizations/community members who raise different concerns with the addition/change of the sex, gender, and/or sexuality questions. I present a few key concerns and some of the resulting best practices below.

Principle of Inclusion: Allowing Self-Identification

There were several initial pushes to change the single, binary sex question, not only for methodological reasons but also to ensure an identity-affirming approach for those who do not identify with their sex assigned at birth. For example, in the rehearsal questions for the 2021 UK census, trans respondents who were asked a sex question were particularly likely to “not consider the question to be acceptable or relevant,” and instead find the question “intrusive” (ONS n.d.). Further, the ONS found that “in every qualitative test, participants who identified as trans stated that it was important not to force individuals to disclose that they are trans, as it may not be safe to do so” (ONS n.d.). This provided suggestions that i) the questions for sex might be difficult for gender minorities, and ii) the questions for certain identities should be made optional.

Data Misclassifications

An overarching tension between data users and queer organizations/community members has been that of data concordance. Many quantitative social scientists expressed concerns that replacing a sex question with a gender question would disturb the ability to compare longitudinally and for equalities monitoring. Nonetheless, other scholars have shown that there is measurement error associated with sex as it has been measured—e.g., by interviewer-coding (Lagos and Compton 2021), and self-identification could reduce error. 

Survey Specific Concerns

Other typical survey specific concerns, such as question ordering, response options, and response type, should be particularly flagged for concern given the sensitivity of identity questions. For instance, the ONS had to test whether to put the sexuality or gender identity question first because of nonresponse issues or survey dropout (ONS n.d.). Ahead of the 2017 census, Stats NZ found that more people refused to answer a sexual orientation question than identified as non-heterosexual; this finding compelled Stats NZ to delay the fielding of a sexuality question since they did not want to introduce more survey nonresponse than LGBTQ+ identity disclosure (Cooley 2019). Further, there are concerns about response types. Stats NZ, for the 2018 survey, had initially requested that intersex people request a paper form to mark both the “male” and “female” boxes on the sex question given the online form would force the validation of only one option (Strongman 2018).

Census Specific Concerns

Beyond broad concerns relating to survey methods and analysis, there are census specific concerns. This includes the primary respondent of censuses (which is often the head of household) or the age that respondents must be before they can answer certain sociodemographic questions. In the UK, for instance, the head of household is often the primary respondent of the census, which means that all in the said household have to be “out” to the head of household. Additionally, in the UK, only those over the age of 16 are allowed to record their sexual or gender identity, though notably LGBT youth have been driving the diversification of sexual orientation disclosure in recent years (ONS 2018; Gallup 2021, 2021).

Ethical reporting standards: Ensure prevalence is not conflated with disclosure

All estimates produced from sexual and gender identity data should only be expressed with caveats of probable undercounting, and all analyses using these estimates should appropriately designate that this is not necessarily a true population estimate (i.e., a measure of disclosure, not a measure of total prevalence). For example, undercounting “could play into the hands of those who would attempt to reverse progress towards equality” (Cooley 2019), since it would potentially provide “population-level” numbers that might be much smaller than actual prevalence (i.e., the “true” parameter); those who are more antagonistic towards LGBTQ+ rights might use these small numbers as a means of justifying LGBTQ+ erasure.

Complexity of queer identities, changing times

Finally, stakeholders have also mentioned concerns about attempting to capture fluid and potentially unstable identities. One of the key methodological critiques about using large-scale surveys (and correspondingly population and aggregate-data methods) to capture concepts like gender and sexuality is the idea that queerness cannot be quantified. Further, putting sexuality and gender into boxes is, for some people, contradictory to the idea of what sexuality and gender are since they are often context and socially dependent (Ventriglio and Bhugra 2019; Deaux and Major 1987). Changing the fundamental way that queerness is asked in a survey (e.g., through different survey tools) may be a helpful way to consider the multidimensionality of queer identity for future census testing. 

Summary

In sum, the process of studying censuses can be fruitful for understanding the concept of measurement more broadly, given the size of the census and the need for methodological rigour. There are several censuses around the world, such as the aforementioned four case study countries, that have undergone several simultaneous alterations on their censuses and/or large-scale surveys to better capture both gender and sexual majorities and minorities. Nonetheless, with all the testing done in each country, several came up with different conclusions to the same problem. In this section, I discussed four countries that have all approached this challenge of data and measurement in different ways, and I have listed some of the corresponding challenges and recommendations from them.