LGBT+ Youth Perspectives on Sexual Orientation and Gender Identity Questions in the Growing Up in Ireland Survey: A Qualitative Study

Abstract

The increasing importance of identifying lesbian, gay, bisexual and transgender (LGBT+) populations is a key driver in changes to demographic data collection in representative surveys of youth. While such population-based data are rare, Growing Up in Ireland (GUI), an Irish, government-funded, longitudinal survey, includes sexual orientation and gender identity (SOGI) measurements. This qualitative study responds to a query from the GUI study team and aims to identify how best to collect SOGI data in future waves of GUI. A university Human Research Ethics Committee granted approval for online consultations with LGBT+ youth (n = 6) with experiential expertise in policy making. The research is underpinned by rights-based public patient involvement (PPI) with recorded discussions, which were transcribed and imported into NVivo 12, generating the theme “recognition in research, policy and society”. This co-created article, with the LGBT+ young PPI Panel members, commends the inclusion of SOGI data in GUI and recommends changes in question placement and phrasing. Aligning with best practice, the PPI members provide a template for wording on consecutive sex and gender questions, expanded sexual orientation identity categories and maintaining the existing well-phrased transgender question from GUI. This offers potential to improve the quality of the SOGI data collected and the experience of those completing the questionnaire. These findings extend beyond GUI, with relevance for surveys with youth populations. This paper underscores the potential and benefits of participatory approaches to research with youth and views their role beyond simply as sources of data.

1. Introduction

This co-created article describes a qualitative study into Growing Up in Ireland (GUI), a national, longitudinal survey, which introduced questions on sexual orientation and gender identity (SOGI) [1,2,3]. It explores how lesbian, gay, bisexual and transgender (LGBT+) youth experience that survey. The study developed in response to a query from the GUI study team related to the challenges in capturing the breadth of diversity within youth identities and orientations and asked about better SOGI question wording for future data collection [4]. It contributes to an emergent research literature, which is driving changes to demographic data collection in population-based studies with youth [5]. This study is part of a larger LGBT+ youth project (LGBT+ YuPP☺ Project), which prioritizes public patient involvement (PPI) with LGBT+ youth with experiential expertise of participatory policy making, of a similar age to GUI Cohort ‘98 [4,6,7,8]. This article reports on just one of three themes: “recognition in research, policy and society”.

1.1. Counting LGBT+ Communities

The LGBT+ acronym encompasses lesbian, gay, bisexual and transgender identification, with the “+” plus extending to expansive sexual and gender identities, including questioning, queer, asexual and non-binary people [9]. The umbrella term comprises sexual orientation and gender identity (SOGI) [5,9,10,11,12,13,14,15,16,17]. Sexual orientation comprises identification, behavior and attraction [5,10,15]. Gender identity includes someone’s internal sense of gender, which may not accord with binary concepts of male/female or birth-assigned sex [12], alongside a potential affiliation with a gender community [5,18]. Everyone has a sexual orientation and gender identity [13,14,15,17]. For example, transgender youth may identify as heterosexual or as a sexual minority [19,20].

Population-based data collection is important, however LGBT+ communities are often not counted [13,14,15,17,21]. As a consequence, estimates of the size of LGBT+ populations vary [10,13]. Growing research attention to improving SOGI demographic data collection with adults has been evident for well over a decade [10], with a special issue in the Journal of Official Statistics [21], culminating in the 2022 NASEM publication Measuring Sex, Gender Identity, and Sexual Orientation [13]. The lack of standardization has led some countries to introduce, or develop, recommendations and guidelines for SOGI measurement [16,22,23,24,25,26,27]. This reflects a dearth of prevalence data for LGBT+ populations, and extends to the Irish context, with a reliance on approximations in the absence of official data collection [28].

1.2. Population-Based SOGI Data Collection with Youth

The focus on adult population-based SOGI measurement likely reflects the limited availability of prevalence data for younger cohorts [13]. Research on inclusive SOGI measurement with representative youth samples is an emergent topic area [5]. Where such data exist, many surveys record a higher prevalence for youth than older age groups [9,29,30,31,32] This suggests greater familiarization with SOGI terminology, with younger generations more likely to have grown up in supportive social environments [21,33,34]. Attention to SOGI measurement with youth coincides with the availability of quantitative population-based datasets in the US, which introduced variables on sexual orientation identification (rather than attraction or behavior), such as the 2015 US Youth Risk Behavior Survey [19,35,36]. Datasets introducing questions on gender identity are more recent again, such as the 2015–2017 California Healthy Kids Survey [37] and the 2015 Minnesota Student Survey [38,39].

Survey methodologies are not value-neutral [40], rather, there are complexities in relation to measurement, with a lack of consensus on how to define and conceptualize SOGI constructs [5,21]. Measures of sexual orientation identification are considered most relevant in determining health inequalities [10,13,16]. However, survey response options for sexual orientation identity are often restricted to heterosexual/straight, gay/lesbian, and bisexual [21,27,41]. Further, rather than regarding gender as a spectrum [9,42,43], sex and gender may be conflated or a single binary measure of male/female used which can lead to “mismeasurement or misuse of the data” [13] (p. 25). Where a separate gender question is asked, non-neutral language, such as ‘other’ may be offered [15,26,44,45]. Inclusion of a specific question on transgender identification is rare; and may mistakenly be included as a sexual orientation category [20], or gender response option [38,39,46]. Further, some youth may not use ‘transgender’ terminology and identify as male or female [5,45,47]. Such issues potentially impacts on representation, with LGBT+ youth unable to participate, or accurately record their SOGI identity, resulting in misclassification and undercounting [9,13,15,19]. As such, there are concerns regarding research validity, with an impact on analysis and reporting, undermining the quality of prevalence estimates [13,14,17]. Alongside such considerations within survey design, there is an equally pressing need for majority populations to understand SOGI questions [13,14,18].

While such population-based data can be rare, it is in this context that GUI is the first Irish representative study to collect data which included SOGI variables.

1.3. Growing up in Ireland (GUI)

GUI, the national longitudinal study of children, is commissioned by the Irish government, focusing on the lives of children and youth [1,48]. Four waves of data have been collected from Cohort ‘98 at 9 years, 13 years, 17–18 years and 20 years [49,50]. SOGI questions were introduced in Wave 3 [1,2]. The data collection spans a unique period of time, with Cohort ‘98 respondents born five years after the decriminalization of homosexuality and coming of age at the time when two-thirds of the Irish population voted for marriage equality [51,52,53]. Cohort ‘98 respondents completed a Main Questionnaire alongside a Sensitive Questionnaire [3]. As Figure 1 illustrates, sex was recorded in X1, alongside other demographic data, with questions on sexual orientation, gender and transgender self-identification included separately in the Sensitive Questionnaire in Section D titled “Gender identity and intimate behaviour” [3].

Attention to privacy and confidentiality was prioritized during data collection, using computer-assisted self-interview, with SOGI responses provided via self-report from the young person [1]. Following archiving, anonymized data are made available on a confidential basis, with statistical disclosure control criteria strictly applied [54,55].

With implications for policy and practice, GUI seeks to ensure that the data collected address identified research needs, maximizing data use [49,50,56].

1.4. Irish LGBTI+ National Youth Strategy and Youth Advisory Group (YAG)

Ten percent of 17–18-year-olds in Ireland identify as lesbian, gay, bisexual, transgender, questioning, asexual, or describe their gender as other (LGBT+) [57,58]. This relatively high prevalence accords with international findings and likely reflects improved social and legal climates [57,58]. Despite social progress, subsequent research suggested substantial vulnerabilities [59], and led to the initiation of the Irish LGBTI+ National Youth Strategy [6,7,52,60]. A Youth Advisory Group (YAG) was core to strategic development; the participants of the PPI Panel are members of this group [52,61]. This study is aligned to Goal 3, Objective 15: Enhance the quality of LGBTI+ research research and to:

• 15(d) Review… appropriate language and ways to ask about gender identity and sexual orientation… to inform the development of best practice instrumentation… for inclusion in surveys and/or Census.
• 15(g) Explore Growing Up in Ireland Wave 3 data that captures sexual orientation and other relevant information [52] (pp. 30–31).

While policy often commissions research, the prioritization of the research and data environment as one of its goals noteworthy [52]. This accords with NASEM who concluded that despite increased knowledge of health disparities, “glaring gaps remain, often driven by a lack of reliable data” [13] (p. 2).

Health inequalities for LGBT+ communities can be understood within a social justice framework, with “recognitive justice” crucial to revaluing disrespected identities [62] (p. 12).

1.5. Participation as a Form of Social Justice

Honneth’s Recognition Theory conceptualizes three intersecting spheres of recognition: among individuals; within laws and policies; and by communities [63,64,65]. This tripartite framework is explicitly extended to advocate for full child and youth participation [66]. The involvement of youth in research is particularly important in light of the commitment within Irish policy, which accords with the UN Convention on the Rights of the Child (UNCRC) [67]. This offers a framework for the many levels that rights might encompass: decision-making; non-discrimination; direction and guidance; and access to information [67,68,69,70]. Despite policy intentions, it has been acknowledged that some young people are less likely to be consulted, and are seldom heard as a result, including LGBT+ youth [71,72]. In response, The Lundy Model of Participation was adopted across Ireland, with an emphasis on the inclusion of seldom-heard youth [71,73,74,75]. Such rights-based approaches align with efforts to ensure meaningful involvement and engagement of patients and members of the public (PPI) in research [69,76,77]. Described as “research being carried out ‘with’ or ‘by’ members of the public rather than ‘to’, ‘about’ or ‘for’ them”, PPI is emphasized as integral to health research [76] (p. 1). It seeks to ensure that research undertaken is relevant [66,69,77,78,79]. Participation may be particularly important for seldom-heard youth [66,68,70,80,81,82,83]

1.6. Rationale

Attending to SOGI placement and phrasing in population-based survey design is timely; better SOGI measurement will improve the ability to identify LGBT+ youth, generate more accurate prevalence estimates and understand the lived experience of these populations [13,14,17]. While the study developed in response to a query regarding GUI [4], it has broader relevance as the first Growing Up study with a representative, longitudinal dataset that includes SOGI variables [1,49,50,56]. This study may be of interest to other Growing Up projects in other jurisdictions: Australia, New Zealand and Scotland. It may also be of broader global interest, given the recent availability of population-based datasets with youth samples, driving interest in this emergent topic area [5].

While cognitive testing is typically used for exploration of survey design [84], this more comprehensive study, recognises the complexity of SOGI measurement, and accords with NASEM recommendations for “expanded testing among youth” [13] (p. 13). PPI is emphasised as integral to research in the health context, increasing effectiveness, strengthening methodological rigor, and enhancing both the quality and relevance of the research [78]. To the best of our knowledge, this study is unique in exploring the views of LGBT+ youth on SOGI question placement and phrasing within quantitative surveys.

2. Materials and Methods

As mentioned previously, this research is part of a larger research project called the LGBT+ YuPP☺ Project, which convened Young People’s Panels to garner youth perspectives and priorities on LGBT+ wellbeing. This larger project is informed by Recognition Theory [63,64,65] and the Lundy Model of Participation [68,70], and is embedded in rights-based PPI [69,76,77,78,79]. This Project is described elsewhere [6]. The PPI participants involved in this study are all involved in the larger project. The Standards for Reporting Qualitative Research (SRQR) were followed [85] (see Appendix A,

Table A1. Standards for Reporting Qualitative Research (SRQR).

No.	Topic	Item	Manuscript Page No.
Title and abstract
S1	Title	Concise description of the nature and topic of the study identifying the study as qualitative or indicating the approach (e.g., ethnography, grounded theory) or data collection methods (e.g., interview, focus group) is recommended	1
S2	Abstract	Summary of key elements of the study using the abstract format of the intended publication; typically includes objective, methods, results, and conclusions	1
Introduction
S3	Problem formulation	Description and significance of the problem/phenomenon studied; review of relevant theory and empirical work; problem statement	2–5
S4	Purpose or research question	Purpose of the study and specific objectives or questions	6
Methods
S5	Qualitative approach and research paradigm	Qualitative approach (e.g., ethnography, grounded theory, case study, phenomenology, narrative research) and guiding theory if appropriate; identifying the research paradigm (e.g., positivist, constructivist/interpretivist) is also recommended	7
S6	Researcher characteristics and reflexivity	Researchers’ characteristics that may influence the research, including personal attributes, qualifications/experience, relationship with participants, assumptions, or presuppositions; potential or actual interaction between researchers’ characteristics and the research questions, approach, methods, results, or transferability	7.
S7	Context	Setting/site and salient contextual factors; rationale	5–6
S8	Sampling strategy	How and why research participants, documents, or events were selected; criteria for deciding when no further sampling was necessary (e.g., sampling saturation); rationale	8
S9	Ethical issues pertaining to human subjects	Documentation of approval by an appropriate ethics review board and participant consent, or explanation for lack thereof; other confidentiality and data security issues	9
S10	Data collection methods	Types of data collected; details of data collection procedures including (as appropriate) start and stop dates of data collection and analysis, iterative process, triangulation of sources/methods, and modification of procedures in response to evolving study findings; rationale	9
S11	Data collection instruments and technologies	Description of instruments (e.g., interview guides, questionnaires) and devices (e.g., audio recorders) used for data collection; if/how the instrument(s) changed over the course of the study	9
S12	Units of study	Number and relevant characteristics of participants, documents, or events included in the study; level of participation (could be reported in results)	8
S13	Data processing	Methods for processing data prior to and during analysis, including transcription, data entry, data management and security, verification of data integrity, data coding, and anonymization / deidentification of excerpts	9–10
S14	Data analysis	Process by which inferences, themes, etc., were identified and developed, including researchers involved in data analysis; usually references a specific paradigm or approach; rationale	9–10
S15	Techniques to enhance trustworthiness	Techniques to enhance trustworthiness and credibility of data analysis (e.g., member checking, audit trail, triangulation); rationale	10
Results/Findings
S16	Synthesis and interpretation	Main findings (e.g., interpretations, inferences, and themes); might include development of a theory or model, or integration with prior research or theory	9–21
S17	Links to empirical data	Evidence (e.g., quotes, field notes, text excerpts, photographs) to substantiate analytic findings	9–21
Discussion
S18	Integration with prior work, implications, transferability, and contribution(s) to the field	Short summary of main findings; explanation of how findings and conclusions connect to, support, elaborate on, or challenge conclusions of earlier scholarship; discussion of scope of application/generalizability; identification of unique contribution(s) to scholarship in a discipline or field	9–21
S19	Limitations	Trustworthiness and limitations of findings	18–21
Other
S20	Conflicts of interest	Potential sources of influence or perceived influence on study conduct and conclusions; how these were managed	25
S21	Funding	Sources of funding and other support; role of funders in data collection, interpretation, and reporting	24

).

2.1. Aim and Objectives

This study explored PPI Panel members insights into SOGI question design in GUI. The overarching aim was to generate suggestions for possible changes to SOGI question placement and phrasing for future GUI data collection. This sought to improve both the data and the experience of the respondents asked to provide those data, enhancing data quality [5,15].

2.2. Researcher Characteristics and Reflexivity

The first author and lead researcher (N.C.) was a member of the Oversight Committee, while the PPI Panel were former YAG members for the Strategy [6,7,52]. In this capacity they were consulted on the wording of SOGI questions for the Strategy’s nationwide consultations and questionnaire, gaining awareness of the challenges in survey design [52]. Drawing from this experience, alongside their lived experience, the PPI Panel was uniquely positioned to provide valuable insights. Their experiential expertize spans the Irish marriage equality referendum, introduction of legal gender recognition by self-identification for adults, SOGI data collection with Cohort ‘98 in GUI, alongside their involvement in the Strategy [28,51,52,53,60]. Insider research is considered beneficial in studies with LGBT+ communities [86]. The “explicit incorporation” of experiential knowledge in research [87] (p. 225), accords with PPI [69,76,77,78,79]. Within ethnographic approaches, researcher subjectivity and positionality are regarded as integral, with an emphasis on reflexivity [8,88,89]. ‘Emic’ and ‘etic’ are used to contrast research participants’ experiential expertise (emic), with researcher knowledge (etic) [88,89]. Problematizing this artificial binary dichotomy, a memo writing approach was adopted by the lead author, providing a reflexive resource [8,87].

2.3. Access, Recruitment and Sampling

The first author and lead researcher attempted to contact all 13 members of the YAG, through the original Strategy network, with all but one member from the original group informed of the YuPP☺ Project. This is described elsewhere [6], and included in Supplementary Materials (see File S1). Information on the research design and aims was shared with the 12 members contacted. While nine of the 12 expressed interest, six LGBT+ youth were recruited, from across Ireland. The PPI Panel held multiple roles: as an advisory group for the YuPP☺ Project; as participants in online consultations; in co-producing research; and as youth co-authors [61,79,80]. At all times, it was emphasized that PPI Panel were representing themselves, and we are unable to comment on the representativeness of the final sample to the 13 original YAG members. As such, the ‘information power’ held within the sample was recognized [90], and the study did seek ‘saturation’ [91].

Their characteristics are outlined in

Table 1. PPI Panel member’s characteristics.

Characteristic	Number of PPI Panel Members
Grew Up in Ireland	yes (n = 6) rural (n = 3), urban (n = 3)
Sexual orientation	lesbian (n = 3) gay (n = 3) queer (n = 1) pansexual (n = 1) *
Gender	male (n = 3); female (n = 3)
Gender identity	transgender (n = 1); cisgender (n = 5);
Pronouns	she/her (n = 3); he/him (n = 3); they/them (n = 1) *
Age at the time of the marriage equality referendum	15 years (n = 3); 17 years (n = 2); 22 years (n = 1)
Age at time of legal gender recognition by self-identification for adults	15 years (n = 2); 16 years (n = 1); 17 years (n = 2); 22 years (n = 1)
Growing Up in Ireland (GUI) study: Aware of GUI  Aware of SOGI data collection in GUI  Participated in GUI  Age at time GUI Wave 3 data collection completed	yes (n = 6) no (n = 6) no (n = 6) 16 years (n = 2); 17 years (n = 1); 18 years (n = 2); 23 years (n = 1)
Age at time of initiation of LGBTI+ National Youth Strategy	16 years (n = 2); 17 years (n = 1); 18 years (n = 2); 23 years (n = 1)

* Due to multiple forms of coexisting identifications, does not sum to six.

.

2.4. Ethical Considerations

Ethical approval was granted by a university Human Research Ethics Committee. Information sheets about the study, outlining the purpose of the research and what participation involved, alongside details of the consultations using online forums, were forwarded to young people who expressed interest in the study (n = 12) (see File S1). In accordance with statistical data control measures, and to prevent identity disclosure, ethical approval specifically stated that no GUI Cohort ‘98 respondents would be recruited. There was no overlap between the PPI participants and the GUI Cohort ‘98 respondents. The voluntary and informed consent of each PPI Panel member was obtained prior to participation (n = 6). A de-identification strategy was adopted through consultation, with unattributed, anonymous quotes presented in tabular format. PPI Panel members were given a nominal gratuity for each stage of the information–consultation–reflection cycle, in acknowledgement of their time and contribution.

2.5. Data Collection

Online qualitative consultations (one-to-one, pair and group) provided data on the survey design and SOGI questionnaire items in GUI. In advance of meetings via ZOOM™, PPI Panel members listened to a Gallup podcast [92], read a newspaper article, [93] and an Op-ed [58]. The first and second author facilitated an online discussion and Q&A on measuring LGBT+ communities with a recognized expert on the geography and demography of LGBT+ populations (see File S1). The first author used the presentation and Q&A to iteratively develop a topic schedule, incorporating the following topics: survey methodology, data collection, confidentiality, questionnaire design, data access, what was measured and how this was measured (see File S1). Between April 2021 and June 2022, four rounds of data collection took place, complemented by two rounds of member reflections. Final feedback sessions were held in September 2022, with the lead researcher (N.C.) sharing findings in the context of best practice [5,10,13,14,15,17,41,44,45,46]. This process of dialogue and feedback, while time consuming, accords with rights-based PPI [67,68,69,70].

2.6. Data Management and Analysis

Online consultations were transcribed verbatim by the lead author, facilitating full immersion in the data. Transcripts were imported into NVivo 12 (QSR International Pty Ltd. Version 12). Codebook thematic analysis offers a pragmatic approach in “meeting predetermined information needs” [94] (p. 7). The data were initially analysed using an inductive open coding technique, with preliminary coding completed twice. Outputs were shared with co-authors, with similar codes grouped together, collapsed, expanded or renamed [94], through an iterative “critical friends” process [95] (p. 113). Themes were generated through moving between the data and codes, with close consideration of the project aims [61]. PPI Panel members actively engaged in this process, providing rich insights and deepening understandings on issues that they felt were important [96,97]. Patterns of broader significance strengthened this process through description and analysis of the rich data [98]. Allowing sufficient time, beyond once-off consultations, or occasional involvement [61,66,69,76,77,78,79,81,82], alongside an emphasis on research processes and relationships [8], resulted in a nuanced and complex consideration of the topic. This reflects the term crystallization, described by Tracy and Hinrichs, where the goal is to develop a more rigorous, complex, multi-dimensional interpretation, enhancing trustworthiness [98], described elsewhere [6]. During a further analysis of the data while writing this paper, with the second author acting as a reflexive partner, it was determined that the results and discussion would be combined in order to contextualize the PPI Panel members’ views within the literature, alongside researcher interpretations [94,99].

3. Results and Discussion

This qualitative data collection project focuses on assessing the utility and accessibility of SOGI data collection in GUI. Results are presented with accompanying discussion, in order to draw connections between PPI Panel members’ reflections, reflexive memo writing, expert consultation, and the extant literature on SOGI measurement. Combining the results and discussion enabled important exploration of contextual experiences (e.g., social change as a backdrop to the GUI survey), and within cultural perceptions (e.g., improved data collection, improved respondent experience), in shaping the PPI Panel members’ perspectives and priorities [94,99]. Situating the joint analysis and interpretation within broader social and cultural contexts recognizes that: “[w]ithout the voice of youth, research can miss the contextual input necessary to represent the unique youth experience” [80] (p. 177). Through codebook analysis, the theme “recognition in research, policy and society” was generated. Figure 2 illustrates the coding concepts and developing categories for this overarching theme.

3.1. Context, Documenting Change over Time and Usefulness of Data

PPI Panel members spoke about the changing social context within which data were collected for GUI, highlighting their awareness of social change resulting in increased LGBT+ acceptance in Ireland, which in turn led to increased LGBT+ visibility [28,33,34,51,52,53,60]. The qualitative narratives highlight an improved social climate for LGBT+ communities within the lifetime of PPI Panel members and GUI respondents [28,33,34,51,52,53,60] (

Table 2. PPI Panel members’ reflections on the social context for the Growing Up in Ireland survey.

Social Context	Quotations to Illustrate PPI Panel Members’ Perspectives and Priorities
Survey Context
Social acceptance	“Really good timing that the Growing Up in Ireland has respondents born after decriminalization and who came of age at a time of marriage equality.So, it’s an interesting time that this data covers”.
	“Identity ebbs and flows and changes all the time. These young people could be in their 70s and be asked the same question and give a different answer, when the Growing Up in Ireland survey turns into the Growing Old in Ireland survey”.
	“As societal acceptance increases and things continue to change, it will be interesting to see how that data changes with this wider change, because it’s longitudinal… Or that with the impact of living in a rural community, we don’t see those changes.”
Valuable data	“They are official and without them people are not represented. So, it is very important to have data.It’s so hard to get that information. I would always be excited about this”.
	“It is very powerful for people – in terms of justifying and explaining their experiences and verifying the needs of the community—people who will give funding, who will give rights; people who have the power to do things that will improve these experiences”.
	“A project like this that’s longitudinal and representative is the gold standard for people to use. Especially for LGBT+ young people about what the world is like at this age, and across different points in time.”

). GUI provides a unique opportunity to track further progress. Half the PPI Panel spent their childhood and adolescence in rural areas, commenting on expectations of shifts across time and geography, with GUI providing important forms of data comparability. Drawing on their experience of participatory policy making for the Strategy, PPI Panel members highlighted the policy implications, emphasizing how valuable the GUI data are. This includes the potential to provide an evidence base for assessing policy efforts to address acknowledged health disparities [56]. The PPI Panel were attuned to the benefits of GUI as a population-based, longitudinal study, describing this as ‘the gold standard’ [13]. This echoes commitments within Irish policy to “ensure that disaggregated data on the LGBTI+ population and their needs is collected in national surveys where relevant” [53] (p. 19). This reflects calls within the international literature for the routine collection of SOGI data within population-based, representative surveys, with both longitudinal and cross-sectional designs, and within the Census [13,14,15,16,21].

3.2. Survey Design

Discussions included research validity—the extent to which concepts are accurately measured in surveys. They considered the survey design in relation to the title and the alignment of questions. As illustrated in

Table 3. PPI Panel members’ reflections on research validity and SOGI survey design.

Survey Design	Quotations to Illustrate PPI Panel Members’ Perspectives and Priorities
Realigning GUI Questions
Survey title	“So, they’ve put down gender and sexuality as a sensitive topic.I wouldn’t see it as being a ‘sensitive’ topic.It gives a different sort of message to the young person, to the researchers, to the parents as well, if it’s ‘sensitive’
	It is far safer to say ‘supplementary’ because if it’s ‘sensitive’ there are connotations of additional needed privacy. If it’s ‘supplementary’ it’s just extra questions”.
Demographic data	“I can’t quite fathom why that’s so separate from general demographic info. I would always put questions like this in a demographic section. My instinct is that’s where they should go. The information is functionally about demographics—in terms of asking if somebody is LGBT+, it’s a demographic question”.
	“It seems out of place, if that makes sense. I nearly think that Section D was to try to find somewhere to put the sexual orientation and gender identity questions in. So yeah, definitely that question order should be re-jigged”.
	The questions about talking openly about sex could easily go into the sex education section. Then you put the questions about boyfriends and girlfriends into the relationship section. It basically leaves the sexual orientation and gender identity questions sitting on their own—and theyshould just be demographic—sothey wouldn’t need a Section D”.
Separating data collection of sex and gender	“One of the things that strikes me that at the beginning of the questionnaire they ask sex with two options: male/female. Then later on in Section D they ask a gender question with options: male/female/other/prefer not to say—when they’ve already asked about the young person’s sex”.
	“Otherwise, it becomes, ‘Don’t you believe me, or did I give you the wrong answer?’ Especially because they didn’t give the option earlier for people to express themselves”.
	“Were they expecting a trans person to put down their assigned gender in response to male/female? And then a few questions later put down their gender identity and whether they were trans? It can’t just be ‘what is your sex?”
Discussing sex with parents	“That’s the order with this question? ‘Do you find it hard to talk about sex with your parents and are you gay?’ One of the biggest worries that young LGBT+ people first have is ‘What will my parents think?’”.
	“I think it would have an impact on how people might answer. It might be very difficult to answer the sexual orientation question. If someone thinks ‘I can’t talk to my father about sex, let alone sexuality and gender, I can’t be out, I’m just going to put something else’.”
	“They don’t ask whether you can talk openly about sexual orientation or gender identity with your parents, it’s specifically about talking openly about sexual intercourse.They have a question on sex education, so those questions would be much better placed in that section, because the primary educator is a parent”.

, PPI Panel members reviewed the description of the survey as a “Sensitive Questionnaire”. They concluded the use of sensitive has a “connotation of additional needed privacy.” This may reflect historic concerns at how SOGI information has been used or shared [15]. With privacy and confidentiality strictly observed within GUI, this may provide reassurance [3,48,54,55]. This accords with recommended strategies to improve SOGI measurement [5,22,100]. Rather than sensitive information, LGBT+ identification is considered a positive aspect of identity (Table 3). Implicit assumptions that LGBT+ populations are less likely to participate in surveys are not supported in the literature [14,101]. PPI Panel members suggested the use of “Supplementary Questionnaire” was more appropriate.

Realigning GUI Questions

SOGI data were consistently and repeatedly described as demographic data by the PPI Panel (Table 3). However, in GUI, SOGI data have been separated from other demographic data, such as sex and date of birth, and are included in Section D labeled “Gender identity and intimate behaviour” (Figure 1). This reflects a wider pattern noted in the literature where SOGI data collection “are often separated from other demographic measures” [14] (p. 1). Discussions with the PPI Panel concluded that SOGI questions are more appropriately placed within demographic data collection, in this case alongside date of birth. They offered suggestions for realigning the questions (Table 3 and Figure 3). These suggestions accord with recommendations in the literature [13,14,18,25,45].

The importance of realigning Section D was perhaps most starkly noted with the separation of the collection of demographic data, with X1 asking about sex, with binary response options of male/female and Section D3 categories: male/female/other/prefer not to say (Figure 1). The PPI Panel worried that the separation of the collection of these demographic data had the potential to negatively impact on Cohort ‘98 respondents, particularly when they were not given an earlier opportunity to self-identify, leading to discomfort (Table 3). Such concerns reflect the emergent literature [26,42,43,44,46]. There was consensus that consecutive sex and gender questions would better align with the collection of demographic data at the beginning of the questionnaire, reflecting best practice [13,14,24,25,44,45,46]. The use of two-step processes, with birth-assigned sex, followed by a gender question with expanded options, beyond a male/female binary, offers the potential for “nuanced, tailored analyses that better reflect unique experiences within this population” [46] (p. 249).

PPI Panel members reviewed the question preceding SOGI questions in relation to discussing sex with parents. They expressed concern that the question order inadvertently suggests an association between sexual behavior and sexual orientation. Section D is described as “Gender identity and intimate behaviour” [3]. This appears to conflate sexual orientation identity with behavior. The PPI Panel felt this question order could negatively impact on data collection, resulting in undercounting [100,102,103]. They described parents as first educators and thought that the questions would better align with Section C on sex education (Figure 3).

3.3. Survey Questions

The PPI Panel also reflected on the phrasing for questions on sexual orientation, gender and gender identity (

Table 4. PPI Panel members’ reflections on research validity and SOGI survey questions.

Survey Questions	Quotations to Illustrate PPI Panel Members’ Perspectives and Priorities
Question Phrasing: Sexual Orientation
Measuring sexual orientation	“Just thinking about identity vs attraction. I know someone who tends to identify as gay, even though attraction-wise they often find themselves attracted to multiple genders… more in line with pansexuality”.
	“They’ve got questions that captured identity, despite descriptions of attraction.I think pretty much for my age group, if they asked: ‘How would you describe your sexual orientation?’ without having the attraction in there, that could work”.
	“So, you’re talking about young people who’ve all grown up with marriage equality as a reference point. It probably did mean that those young people could understand the terms.Even if you’re someone who doesn’t understand ‘heterosexual’—I’m sure my age group would know ‘straight’”.
	“I don’t like that ‘heterosexual’ is first and gay is second, because, again it’s arbitrary, clearly, but it’s also not and implies a hierarchy. I am very taken aback by the fact that they have used heterosexual / straight… There’s clearly a difference”.
Descriptors of attraction	“I’m wondering why does heterosexual/straight have ‘sexually attracted to the opposite sex’, but the rest just have attracted?” That’s a bit leading, it’s like: “So I have to be straight to be sexually attracted, whereas if I’m gay or lesbian, it’s just attracted.
	[Reading responses] heterosexual: sexually attracted to the opposite sex; gay or lesbian: attracted to the same sex. Well, that’s a really good point. It’s definitely an oversight.I would highly doubt that was something that was purposely done”.
	I suppose the asexual part, being a spectrum, to just ask about attraction, is not quite right. I’m pretty sure that there are asexual people that are attracted to people, that are in relationships. I just think the part in brackets ‘not attracted to either sex’, that may not be accurate”.
Inclusion of pansexual	“I think pansexual should be included. I mean if we’re really being inclusive. I have heard people saying: ‘Well I’m not bisexual, I’m pansexual,’ or vice versa”.
	“I might have put it as bisexual/pansexual because there are so many people who see a relatively wavy dotted line between them in their own experience. But I don’t think it would be a bad thing for it to be included separately either.” “With the question about sexual orientation, the bisexual part would assume two genders, so maybe the introduction of pansexual within the survey needs to be added. Even if people are like: ‘Does that mean you’re attracted to pans, ha ha“.
Inclusion of queer	“ I don’t know how much of a risk it would be putting in queer. I wouldn’t want someone to be confronted in the middle of a survey, when it’s not necessary”.
	“I love it. It’s a very valuable term that a lot of people really love and love to use and there are a proportion of people who are upset by it.”
	“I have a very awkward relationship with the term ‘queer’. I didn’t see anything positive about it. Now, living as an adult in Dublin, it doesn’t have those connotations. I love the principle of it, I love that it includes everyone”.
	“My perspective used to be ‘Well, I’m reclaiming the word, it’s my identity, it’s who I am and I’m entitled to use it’. So queer is an identity that I use, but I wouldn’t impose it. You won’t really find anyone who has a neutral reaction to the word queer—you either find it empowering, or it’s upsetting. There is no in-between”.
	“Because that word was used against people growing up, as a slur, not everyone’s comfortable with it. My Dublin-based friends would have no problem; my rural friends would have a much harsher view on the word ‘queer’”
Question phrasing: gender identity
Gender “other” as a response category	“There’s no way to capture what ‘other’ is, because like sexuality, gender is a spectrum. So, by just having binary terms and then ‘other’ it doesn’t allow people to put down what their gender is. It could be that someone is gender fluid, and they don’t identify with gender ‘other’. Maybe gender ‘other’ is assumed to be non-binary, so then that option isn’t there for them to self-identify”.
	“We’re not at a stage yet in Irish society where someone would like casually say ‘gender other’. It’s a lot more normalized to have fluidity in sexuality than it would be in gender. I don’t know whether, in this circumstance, that could be very offensive or feel like erasure to categorise a number of identities under ‘other,’ which is kind of problematic”.
	“It’s hard to probe into what people mean when they say ‘gender other’ because everybody who’s non-binary, they don’t all just say ‘non binary’—there’s a lot of more specific non-binary identity labels that they may use—agender or gender fluid or gender queer are ones that are more common”.
Transgender identification	“Having a separate question about whether or not somebody is trans is the best option. So often, the alternative is just putting trans in the list of sexualities”. When there’s a list of gender options and it’s male / female / trans male / trans female—that’s not good”.
	“I will say points for the ‘Would you describe yourself as transgender’?”

). They emphasized that the continued inclusion of SOGI questions in further waves across the lifespan of GUI may be particularly important in light of the fluidity of identity [16,47]. It is noteworthy that GUI repeated the SOGI questions at Wave 4 when Cohort ‘98 were aged 20 [50], and will include SOGI questions in the forthcoming Wave 5 of data collection with Cohort ‘98 at 25 years [56]. It has been acknowledged that, for youth who may be marginalized, it is important that they “see themselves reflected in the questions” [56] (p. 25).

3.3.1. Sexual Orientation Questionnaire Items

The inclusion of a question on sexual orientation and the range of available responses within GUI, particularly beyond lesbian, gay and bisexual, to include questioning and asexual responses, is commendable (Table 4 and Figure 3) [9,13,21,41,44]. The questioning response is recommended for youth population, at a time when they may be exploring their identities [5,9,13,14,58]. GUI focus groups with 25-year-olds, consulted in relation to Wave 5 data collection, emphasized the importance of SOGI data collection including “anyone who was questioning their sexuality” [56] (p. 75).

The PPI Panel members have a nuanced understanding that sexual orientation is comprised of identity, behaviour and attraction [10,15,104]. This accords with Guyan who notes that “how people understand and define their sexual orientation is not always simple” [16] (p. 786). An example was provided of someone who ‘tends to identify as gay’ although ‘their attraction would be more in line with pansexuality’ (Table 4). However, they observed that sexual orientation in GUI uses attraction in clarifying language [3].

There were contrasting views on the use of the term “straight”, with the response order potentially creating “a hierarchy” (Table 4 and Figure 4) [13]. It is noted a heterosexual/straight sexual orientation was described in relation to sexual attraction, with other orientations described solely in relation to attraction (Figure 4). The PPI Panel members observations are consistent with best practice, which suggests that SOGI data collection should not conflate identity with attraction or identity with behavior [13,14].

While Spock et al. recommend asking about all three aspects of sexual orientation: identification, attraction and behaviour [5], measuring attraction and behavior pose additional challenges, requiring further research [4,10]. A limitation of this study is that question wording on attraction and behavior were beyond the scope of these consultations.

The inclusion of the asexual option, a rarity in population-based data, may help improve data quality [9,13,14,41,44], however, the asexual response in GUI is described in brackets as “not attracted to either sex” (Figure 2). The PPI Panel members challenged the accuracy of this description [13,44,105,106,107]. This may be particularly important as it can be assumed that young people, universally, experience sexual attraction [105,107]. There is a pressing need for measures which recognize broader conceptualizations of attraction, and the absence of attraction, in response categories [13].

Discussions also focused on categories missing from the available range of responses (Figure 4). They suggested that the inclusion of pansexual, a non-binary sexual orientation, would be optimal [9,22,41]. There was further discussion about the term “queer”, which was not included as a category in GUI. PPI Panel members spoke positively about the term “queer” as facilitating broader, more fluid forms of identification. This reflects the first Census 2021 estimates for England and Wales, which facilitated a write-in response, where pansexual and queer were recorded by respondents [108]. This is also consistent with Gallup Daily Tracking data, the largest random, representative sample of LGBT adults in the United States, where respondents can record their preferred sexual orientation [29].

It was also noted that queer may be used in relation to gender, with these questionnaire items discussed below.

3.3.2. Gender and Gender Identity Questionnaire Items

The PPI Panel discussed the gender response options (

Table 5. PPI Panel members’ reflections on “learning with” young people regarding SOGI question placement and phrasing.

“Learning with” Young People	Quotations to Illustrate PPI Panel Members’ Perspectives and Priorities
Enhancing Data Quality
Input into research teams	“I think it’s really good that Growing Up in Ireland are asking about this, because it’s all this language that is really binary, and things are much more beyond the binary”.
	“It is so important to have input into the research teams, to have that kind of objectiveness and have representation from different communities”.
	“I’m glad they’re thinking about how to ask these questions. It’s great that it has been identified as an important thing to look at when designing these surveys”.
Understanding of SOGI questions by the majority youth population	“Young people generally know what their sexuality is…So, to not understand the question would be unlikely. Most people will read a question asking them their gender and they will know the answer.People do not know they have a gender identity and will say: ‘Yeah, I don’t have a gender identity, yet’. Yes, yes you do.They also have no idea if they are trans or not”.
	“So, everybody has different understandings of a question. Even if it’s clear to the person coming up with the question, it’s not always clear to the people answering the question, you know. We have to make sure everyone is included. It’s just that there are people who don’t know what some of the questions mean”.
Inclusion through free-text options	“In an ideal world we would use open-ended questions. It’s great for people answering the questions to feel like they’re being validated. But, you can’t label what someone is saying, you can’t assign them that identity”.
	“I mean it would probably be great to have an open-ended, qualitative answer box, you know. If we don’t have language, I mean documented language, on all the varied and intricate gender identities, maybe it might be a nice opportunity to say ‘Can you please name your gender identity?’”
	“When it comes down to actually crunching figures and using the data, open ended questions are very unhelpful.When it comes to putting that data to use, it’s not as helpful, and I think that that should be the priority. But we need to find a good balance.”
	“If you’re writing in a response, there’s no visibility. If they’re not out to themselves, you’re still creating some kind of visibility for them, for the first time ever, for the first time in a while, or for potentially what they could maybe tick in the future. To have visibility they can see all these options. I think that’s a positive thing”.

and Figure 4). There was a broad observation that “gender other” does not facilitate broader forms of self-identification [15,26,44,45,47]. The lack of a write-in response option was noted which did not facilitate self-expression of fluid forms of gender [5,9,44,45,47]. Such reflections are consistent with Spock et al., who suggest response options should offer a range of gender identities [5]. They expressed concern as to whether the use of “other” was literally a form of “othering”, wondering whether this could be offensive or a form of erasure [15,26,44,45]. This is consistent with NASEM, who call for the use of neutral language that does not negate sexual and gender minority identities [13,14].

The GUI phrasing in relation to transgender identification was commended, highlighting the distinction that GUI provides [3]. This complements the inclusion of a two-step gender question and acknowledges that some people may not describe themselves as “transgender” and identify as male, female or non-binary [5,9,45,47]. The PPI Panel spoke positively about the phrasing of the question (Figure 1). While it was noted that young people may not fully understand the term transgender [18], they would know whether or not they describe themselves in this way.

3.4. “Learning with” Young People

PPI Panel members reflected on the benefits of collaborative explorations of survey context, survey design and survey questions, As a PPI Panel, consideration was given to the understanding of SOGI questions by the majority youth population and the potential of inclusion through free-text options. The PPI Panel welcomed the opportunity to share their perspectives and priorities. This suggests that reflection on, and evaluation of, SOGI question placement and phrasing may be particularly important for these seldom-heard youth. Notwithstanding the exemplary inclusion of SOGI questionnaire items, the importance of the majority youth population understanding SOGI questions was highlighted (Table 5).

Self-identification is highly important, and youth, in particular, more familiarity with, and use a broad range of diverse forms of identification, or reject such labels entirely [4,6,21,44,45,47,109,110,111]. With the marriage equality referendum providing a reference point, the PPI Panel suggested that SOGI questions would likely be understood by Cohort ‘98 respondents, enhanced by the well-phrased transgender question [3]. This is confirmed by prevalence findings, using this data [57,58]. Drawing on their direct experience of the challenges they encountered in contributing to the nationwide survey design for the Strategy [52], PPI Panel members emphasized the importance of inclusion for sexual and gender minority populations and understanding by majority youth populations, recommending further piloting.

Much consideration was given to free-text response options (Table 5). While self-identification is optimal, it is acknowledged by the PPI Panel that this may be less useful for data analysis. They also concluded that visibility in the response options may have a positive impact [5,21,41]. This accords with Honneth, where visibility is a form of recognition [63,64,65]. This is further emphasized with the addition of a response option: ‘I use another term,’ which accords with recommended terminology in other jurisdictions [22].

While this enables respondents to record non-binary gender identities, it is emphasized that this is not a proxy to gather intersex variation [13,14]. A limitation of this research is that intersex status was not discussed, although stakeholder consultations specifically mentioned the need for appropriate question wording for inclusion of non-binary and intersex youth [4]. The needs and context for intersex youth may not be encompassed by terms related to sexual and gender minority populations [112]. As this is not measured in GUI, it was beyond the scope of the study [3,48]. However, this highlights a pressing need for prioritization of research attention for youth with variations in sex development [112,113].

3.5. Summary

While acknowledging the nuance and complexity of SOGI measurement, these rich qualitative narratives highlight subtle, potentially impactful changes to question placement and phrasing. These observations accord with best-practice guidance in the literature, with broader resonance beyond GUI. This highlights the contribution of GUI, enhanced by this co-created analysis, to the emergent literature with youth populations. Fundamentally, it is emphasized that SOGI data are demographic data and are best placed with other demographic questions [13,14,18,25,45].Questions about sex and gender should be asked consecutively, first asking birth-assigned sex, on the original birth certificate, followed by a gender question with the inclusion of a free text option This accords with recommendations for a two-step process [13,45,46,114,115]. This question format is consistent with recommendations from within gender minority communities [115]. A write-in response category with neutral language, ‘I use another term,’ is regarded as inclusive. This conceptualizes gender beyond a binary option of male/female [13,22,26,45,46,47,114].

Sexual orientation self-identification is most meaningful in seeking to identify sexual minority youth experiencing differential treatment, with resultant inequalities in outcome [13,14,17,22,25]. There is the potential to enhance sexual orientation data collection through the expansion of response options, such as pansexual and queer, alongside the inclusion of a write-in response option [9,13,17,22,41,44]. This is supported by prevalence data from the US Gallop survey and the recent UK Census [29,108]. The inclusion of a free text option, would facilitate self-identification [44]. A write-in option may be particularly important for younger cohorts, who use a broad range of self-descriptors [44,45,47,109,111].

Phrasing of the question “Would you describe yourself as transgender?” is commended, with the potential for a slight modification: “Do you describe yourself as transgender?”

Aligning with best practice, Figure 5 provides a template for consecutive sex and gender questions, expanded sexual orientation identity categories and the well-phrased transgender question from GUI.

The template is provided in the spirit of contributing to the emergent literature and advancing the field, and piloting is recommended. It appears that younger people have a greater understanding of SOGI terminology than older cohorts [5,9,18]. For example, it may be unnecessary to list “heterosexual” as the first sexual orientation response, inadvertently creating a hierarchy, or to use “straight” alongside heterosexual [13,14]. Cognitive testing among majority youth populations would be useful in this regard. Further, attending to cultural and other forms of diversity may be important regarding broader understandings of SOGI terminology [13,14,18]. The authors recognize the limitations in the representation of diverse identities and voices within the PPI Panel. This reflects the expressed concerns of the PPI Panel regarding the lack of ethnic and cultural diversity on the YAG [7]. At all times, it was emphasized that the young people were representing themselves and their own ideas, and that these would necessarily be subjective and partial [7]. In particular, piloting is recommended with indigenous and ethnic minority populations, where fluidity and blurring of multi-faceted identities may not captured by the LGBT+ acronym [18]. While acknowledging the limitations of this study, the collaboration with the PPI Panel exemplifies Honneth’s Recognition Theory [63,64,65,66]. This reflects an emergent literature recommending collaborations between researchers and study communities in which a “queer approach to data collection showcases the back-and-forth between participants and researchers.” [15] (p.58). This study may be of interest to other international cohort studies and generate interest globally, particularly within LGBT+ research communities in relation to SOGI data collection with youth and adult populations.

4. Conclusions

While population-based data designed to measure the full LGBT+ community remain limited, GUI is currently the only Irish longitudinal, population-based, government-sponsored survey (regardless of age) that includes SOGI data collection. GUI is exemplary for its inclusion of the following: sexual orientation response categories beyond heterosexual, lesbian/gay and bisexual; separate questions on sex and gender, with gender response options beyond a male/female binary; and the inclusion of a well-phrased transgender question. A welcome request for improved question placement and phrasing of these questions offered a unique opportunity to undertake this study. To the best of the authors’ knowledge, this study is the first to explore the critical insights of LGBT+ youth in relation to SOGI data collection. These rich, nuanced accounts accord with best-practice recommendations. This attests to the social and cultural capital embedded within youth networks and the benefit of “learning with” LGBT+ youth. It emphasizes the importance of rights-based PPI in enhancing “recognition in research”. This accords with Recognition Theory and underscores the “information power” in these valuable narratives, highlighting the quality and relevance of these accounts.

In relation to “recognition in policy”, addressing the lack of nationally representative data, which can be generalized to findings from smaller samples, is a priority. While the collection of official statistics, through the Census, is a stated Irish policy objective, it is with some concern that SOGI data collection in the Irish census is for consideration only, with no clear commitment to implementation. It also appears that the Strategy objective regarding a review of best practice, regarding SOGI data for inclusion in surveys and/or the Census, has not been commissioned. The routine inclusion of SOGI questions and their placement offers the potential to develop standardized forms of data collection, ensuring conceptual precision and enhancing the comparability of studies. This is a matter of some urgency, with far-reaching policy and practice implications.

However, the current placement and phrasing of the SOGI questions within GUI may inadvertently regard these identities and orientations with particular sensitivity. This may reflect wider societal views and may unintentionally create the impression that LGBT+ identities and orientations should be concealed. Self-identification is highlighted throughout the narratives, with broader resonance, given acknowledged health inequalities for sexual and gender minority communities. The importance of counting LGBT+ populations ensures that these marginalized communities count in policy and in society. “Recognition in society” is integral to promoting the social acceptance of LGBT+ youth. While capturing the breadth of diversity within sexual orientation and gender identities poses challenges for researchers, consultations with LGBT+ youth highlight the potential of creative, participatory approaches. The suggested improvements to questionnaire design have the potential to positively impact future SOGI measurement with representative samples. Fundamentally, this will improve the ability to identify and understand LGBT+ youth’s lived experience and inform future research, policy and practice.