A Critical Response to the UK's ‘Sullivan Review’ Into Sex and Gender in Research and Data

1 The ‘Sullivan Review’ in Context

In October 2023, Conservative MP Michelle Donelan, then UK Secretary of State for Science, Innovation and Technology announced, in the context of her desire to ‘[kick] woke ideology out of science’, a ‘review into the use of sex and gender questions in scientific research and statistics’ to ‘[safeguard] scientific research from the denial of biology’ (see Rabesandratana 2023). In response, over 2000 scientists signed a letter exclaiming that ‘[f]ar from depoliticising science, this policy appears to be driven by ideology’ (Rabesandratana 2023, n.p.). The review's terms of reference later specified identifying barriers to rigorous data collection on sex and gender identity within public institutions and broader research systems and developing best-practice guidelines for data collection in these areas (Department for Science, Innovation and Technology 2024). The independent review, commonly termed the ‘Sullivan Review’ after its lead author, sociologist Alice Sullivan, aims to assist government as a funder and producer of research and statistics, alongside other public bodies including universities, research organisations and the health service.

In March 2025, the review's first report, Review of data, statistics and research on sex and gender (Sullivan 2025a; hereafter SR1), was published under the now-Labour government. Its recommendations included that a ‘biological’ conception of ‘natal sex’ should be the default mechanism for recording sex across government and quasi-governmentally commissioned research and data collection; only two, binary options for ‘biological’ sex should be offered (male and female); collecting data around what SR1 calls ‘transgender identities’ should only take place to supplement this approach; the term ‘gender’ should be avoided in question wording; there should be no additional category for people who are intersex/have variations of sex characteristics (I/VSC);¹ and that the NHS should discontinue assigning new NHS numbers and altering gender markers in alignment with individuals’ expressed gender. In July 2025, the review's second report, Barriers to research on sex and gender (Sullivan 2025b; hereafter SR2), was published (although at the time of writing only by the sullivanreview.uk/ website and not the UK Government). SR2 presents adherents to a binary model of sex as possessing ‘ordinary mainstream views about the reality of sex’ (SR2, 32) that align with ‘scientific and scholarly norms which are essential to the production of knowledge’ (SR2, 12). SR2 construes ‘some aspects of gender-identity theory [as lying] outside of truth-seeking norms’, and it describes ‘threat[s] to the norms of the reality-based community’—which it contrasts with the ‘norms of propaganda and disinformation’—as a ‘threat to democracy’ (SR, 31). SR2 also recommends universities consider ‘paring back the scope and power of ethics committees’ in various ways (SR2, 16–17).

In selecting Sullivan, the commissioners chose an advisory group member of Sex Matters, which understands sex as binary, ‘real, immutable and important’ (what it calls ‘objective reality’) and claims, in contrast, ‘gender identity’ as a determination of whether ‘someone is male or female (or both or neither)’ is merely a ‘strongly held and erroneous personal feeling’ (Sex Matters 2024). Prior to selection, Sullivan had made clear she disagreed with research not conceptualising sex as binary and immutable (e.g., Sullivan 2020). While SR1 does not use the phrase ‘gender critical’ to describe its framework, it is allied—through its lead author, terminologies and recommendations—with those self-identifying as ‘gender critical’ feminists. It includes, for example, terms such as ‘adult human female’ (SR1, 11, 43) which some scholars have recognised as an ‘anti-transgender dogwhistle’ (Duffy 2025). ‘Gender critical’ feminism has been analysed by other feminist and queer scholars as part of a broader anti-gender movement (Butler 2025), gender conservatism (S. Ahmed 2021) and a manifestation of ‘postfascist feminism’ (Bassi and LaFleur 2022).² These scholars have demonstrated that this movement's arguments are, as in Bassi and LaFleur's (2022, 317) words, ‘genealogically coherent with multiple conservative moral panics and resilient fascist tropes’. Meanwhile, the Lemkin Institute for Genocide Prevention (2025a, n.p.) has stated that ‘the beliefs held by the gender critical movement overlap significantly with those held by various authoritarian governments and fascist political parties the world over’.

Following Amery (2025), we view ‘gender critical’ projects as biopolitical, as their logics seek what they see as cisgender women's ‘liberation from gender’ through the control, surveillance and restraint of trans and gender diverse people whom they construe as gender deviants. Furthermore, the categories used to collect data carry multiple epistemological and political consequences, and this is acutely the case when individuals and collectives are categorised in projects bearing on the state and its functions. Igo (2008) uncovers how composite statistics shape individuals who contribute to them, arguing that we constitute ourselves relative to who we ‘see’ in surveys. Surveys are fundamental in shaping who is understood as part of the nation/polity and who is not. That social statistics can additionally facilitate individuals ‘imagin[ing] themselves into new collectivities’ (ibid., 285) means categories have significant power in relation to minoritised people. The Sullivan Review's logic attempts to tightly control the potentialities Igo outlines.

Here, we argue that SR1 and SR2 problematically rationalise the existence of trans and gender diverse people as antithetical to truth and scientific legitimacy. We can find no evidence trans or gender diverse people were involved in SR1's design, analysis or writing.³ We also see SR1/SR2 as attempting to undermine critical inquiry that analyses the instability and intersectional production of social categories including sex, gender and sexuality, by questioning the scientific validity and clarity of such study and by arguing this often ‘lie[s] outside of truth-seeking norms' (SR2, 31). We argue that this carries deleterious consequences by foreclosing the ability of social scientific research to analyse social (and biological) processes and by ideologically refusing well-established research practices. Our analysis therefore diverges from the statement by Wes Streeting MP (UK Secretary of State for Health and Social Care) that SR1 ‘will lead to better, more inclusive and fairer outcomes for everyone, including the trans community’ (Hansard 2025).

Trans and gender diverse people in the United Kingdom have had their lived realities denied or erased through still-dominant binarised systems of sex/gender across institutions, resulting in their avoidance of and exclusion from such systems and spaces (e.g., Bonner-Thompson et al. 2021; Lanfear 2021; Todd 2021, 2023, 2024). The largest survey on LGBTQ+ people's experiences in the UK to date, undertaken in 2017, found that 66.5% of trans respondents (n = 14,320) avoided being open about their gender for fear of a negative reaction from others; 75.3%, 58.4% and 55.3% for non-binary respondents (n = 7410), trans women (n = 3740) and trans men (n = 3170), respectively (Government Equalities Office 2019).⁴ Nevertheless, the Sullivan Review does not prioritise furthering trans people's inclusion in research/data. Its recommendations regarding state practices of data collection take place at a moment wherein the administrative flexibility of the state, and of state-regulated functions (including research), are more broadly under attack (e.g., Currah 2025).

We argue that the Sullivan Review contributes to UK ‘anti-trans discourse’ (UN 2024, 6) and ‘eliminationist attacks’ that attempt to ‘criminalise trans existence, remove trans and gender diverse people from public life, and refuse even the reality of trans life’ (Todd 2025, n.p.). These include the UK Equality and Human Rights Commission's (EHRC) interpretations of the Supreme Court judgement in For Women Scotland v The Scottish Ministers, which defined sex in the Equality Act 2010 as ‘biological sex’. The Council of Europe Commissioner for Human Rights (O'Flaherty, 2025, 2) wrote in October 2025 that proposals by the EHRC following that judgement on access to services and facilities ‘have tended towards the exclusion of trans people’, adding that proposed policies could ‘require trans people to habitually “out” themselves publicly’. We join other experts in arguing that the EHRC's subsequent responses will, if unchanged from their draft form and approved by the UK Parliament, erode civil rights and exclude trans people within many public spaces (e.g., Amnesty International 2025; Good Law Project 2025; Lemkin Institute for Genocide Prevention 2025b). Indeed, the Lemkin Institute for Genocide Prevention (2025b, n.p.) has stated that ‘there is a transparent attempt to eradicate transgender and intersex people from British life … a clear example of the 9^th Pattern of Genocide: Denial and/or Prevention of Identity’.

2 Binarisation and Immutable Biology

SR1 recommends using binary terms such as ‘women, men, boys and girls … [as] synonymous with (respectively) adult human females and males and children of each sex’ when describing collected data (SR1, 11). Any deviations from these binary norms of sex at birth—rendered as ‘[s]ensitivities which may apply when referring to specific individuals’—‘should not apply at the aggregate level’ (ibid.). Here, SR1 argues that descriptions of aggregated datasets/groups of people should not acknowledge transness or any sexed/gendered embodiments beyond a binary norm; possibilities obscured in any case by SR1's argument that ‘[t]he default target of any sex question should be sex (in other words, biological sex, natal sex, sex at birth)’, with only binary responses possible (SR1, 5).

SR1 acknowledges arguments ‘against treating sex as a binary variable’ but argues that these ‘include arguments which instrumentalise the lives of people with Differences of Sex development (DSD) and propagate myths’ about the prevalence of these conditions or that ‘people with DSD do not have a sex’ (SR1, 3). Aside from itself deploying people who are I/VSC to make claims about how sex should be conceptualised, the argument that sex is universally determined ‘at birth’ is, again, not accepted by all life scientists and indeed sex characteristics vary and change over the life course for all people, whether transgender, cisgender or I/VSC (Panagides and Keim-Malpass 2025). SR1 (7) also recommends that researchers do not use ‘sex assigned at birth’ to explain sex to respondents. It advocates using ‘sex at birth’ instead, arguing that ‘sex is not determined by [a] birth certificate, it is merely recorded on [it]’, despite also recognising ‘very rare cases an infant's sex may be inaccurately recorded at birth’ (ibid.). In contrast to a sex contextualist approach that accounts for material factors, research context and the intersection of multiple physiologies (Massa et al. 2023; Richardson 2022), these statements negate complexities of sex development, and how both trans and cisgender people's sex characteristics and the behaviour of cells can change through medical intervention and treatments (Massa et al. 2023; Richardson 2022; Southgate and other signatories 2025). Scientists have also argued in favour of more expansive sex conceptualisations that can represent all people to avoid clinical harm, risk in medical decision-making or poor health outcomes. For example, Panagides and Keim-Malpass (2025, 9) have called for medical algorithms to better select specific sexual characteristics of relevance to predicted outcome variables, since ‘[s]ources of allocative harm stem from the inappropriate distribution of resources driven by faulty assumptions that observed group differences are the result of true biological processes’, such as the binarisation of sex.

While SR1 does not deny gender diversity, it construes ‘gender-diverse identities’ through its category of ‘gender identity’, which it maintains is a question of ‘individual psychology’ (19) and of ‘how individuals understand themselves’ (18). ‘Biology’ comes to secure what SR2 describes as sex as a ‘material reality’ (23), envisaged as the appropriate focus of ‘accurate’ data collection in scientific research, in contrast to what is cast as an epiphenomenal sphere of individual self-understanding (which can be disregarded in the practice of science). Through its rigid division between (a certain model of) biology and (a certain model of) psychology, SR1 erases the embodied reality of anyone whose experiences and inhabitation of sex and gender do not conform to that binary. We maintain this constitutes a mechanism of trans erasure and a denial of medical subtleties and embodied lived realities. Indeed, as social scientists, we know that terms like ‘biological sex’, particularly when presented as unquestionable scientific truths and ‘ordinary mainstream views’ (SR2, 32), cannot be seen as apolitical. Calls to depoliticise research and science are themselves ideological.

3 Promotion of Trans Erasure as Clarity, and Deviation From Ethical Principles

How sex and gender are classified and regulated is central to the deepening or ameliorating of inequalities (see Currah 2022), with significant consequences for university communities in particular: SR2 (15–16), for example, pushes against what it calls ‘EDI [Equality, Diversity and Inclusion] overreach’, arguing that LGBTQ+ networks should not be embedded within university and EDI structures and that ‘University EDI departments should not issue guidance on research or data collection’. SR2 (16) also recommends that ‘Universities should consider paring back the scope and power of ethics committees’, delinking committees from university EDI bodies and professionals and advocating altering committees’ terms of reference to state they have no remit to ‘prevent researchers from using the concept of sex, collecting data on sex, or using sex-based language’. SR2 (16–17) also argues that there should be no default formal ethics review of ‘low-risk research which does not involve any medical or other experimental intervention or contact with children or vulnerable populations or genuine risk of harm … or potentially unlawful acts’ (although it does not define ‘low-risk’ or ‘vulnerable populations’, leaving gaps for others to interpret). It recommends that the ‘expansion’ of ethical review to research done by undergraduate students and research comprising secondary data analysis ‘should be halted’ (SR2, 17).

Such notions of ‘ethics’ and ‘clarity’ dismiss lived and embodied experience in research design and deviate from well-established ethical standards. There will be cases where referring to someone as ‘male’ or ‘female’ against their will or knowledge, including when reporting findings on gender or sex at individual and aggregate levels, would constitute misgendering and undermine participant autonomy. This would, we argue, contravene ethical guidance of bodies such as the Economic and Social Research Council (ESRC), which states the primary role of Research Ethics Committees is to ‘ensure that the research will respect the dignity, rights, welfare and, where possible, the autonomy of participants and all parties involved in and potentially affected by the research’ (UKRI: ESRC 2025), and the UK Research Integrity Office principle that the ‘dignity, rights, safety, and wellbeing of participants must be the primary consideration in any research study’ (Parry et al. 2025, 15). Indeed, SR2 (17) explicitly calls for the ESRC to review its ethical guidance, including by acknowledging ‘the risk of politicisation of ethics processes’, a risk which it implicitly regards itself as outside of. SR1 (3) argues that ‘the purpose of survey data collection is to gather data about populations rather than to provide an opportunity for each individual to express the full complexity and richness of their identity’. While we acknowledge that no survey could capture such richness, this statement, taken alongside SR1's data categorisation and university ethics reform recommendations, risks deviating from the ESRC's principle that ‘[t]he rights and dignity of individuals and groups should be respected’ (UKRI: ESRC 2025, n.p.).

SR1 also advocates an ‘observed sex’ approach to recording sex in ‘some face-to-face contexts’ (SR1, 8)—that is, recording a person's sex based on subjective observation. Doing so, SR1 argues, may avoid compromising rapport or ‘exacerbat[ing] a potentially fractious situation’ such as in policing; no further details are offered (ibid.). Rather than breaking rapport, we argue that in the vast majority of situations, offering opportunities for participants to describe/record their own gender or sex constitutes best methodological practice (e.g., Vincent 2018), both to recognise and enhance participant autonomy and agency and to avoid recorders' uncertainty in how to record. We contend that refusing to acquire a person's sex/gender from that person, or removing autonomy in how someone describes their own sex, would constitute an instance where—at the least—their dignity, rights, autonomy and arguably their welfare were not respected (see Todd 2024). Such practices also feed into wider cisnormative and prejudicial socio-political discourse and a hierarchical system of oppression that (a) presents trans people as having mere pathologised gender ‘identities’ which can be easily denied or denigrated, as opposed to the sex and gender of cisgender people which are unquestioned, more highly valued and viewed as more legitimate and ‘natural’ (Lennon and Mistler 2014; Serano 2016) and (b) simplifies a person's sex and gender as always legible to others in line with how they themselves understand and live it. This is troubling given the Review's adherence to a singular ‘birth sex’ model.

SR1 also devalues community knowledges in formulating question design, cautioning that ‘[c]ommunity and campaigning groups should not be assumed to have expertise in question design’ (SR1, 12). While communities may not always have methodological expertise, such statements might undermine researchers’ and survey designers’ willingness to engage with communities. The report also contends that ‘[q]uestions intended for general use should be tested on the wider relevant population, not just on minority groups, including in the case of questions which are primarily intended to identify minority groups’ (ibid.). This argues that minoritised people should only have a limited voice even in aspects of data collection that are likely to be solely about identifying them and their perspectives.

4 Conclusions: Beware of Calls for Clarity

The stakes surrounding the ‘Sullivan Review’ are high given the worsening eliminationist movement against trans and gender diverse people in the United Kingdom and internationally (e.g., Lemkin Institute for Genocide Prevention 2023, 2025a, 2025b, 2025c). The Review adds ballast to attempts to embed a certain conception of biological sex and to separate sex and gender, which it presents as unquestionable science and ‘ordinary mainstream views’ (SR2, 32). It (re)installs ideologically driven models of sex and a hygienically patrolled division between ‘sex’ and ‘gender’, where a certain model of sex is ‘reality’, and ‘gender identity’ is cast merely as an individual's (psychological) ‘self-understanding’—a model that research across disciplines has done much to dislodge (Hemmings 2022; Sothern and Dyck 2009). How the Review does so undermines and discredits critical scholarship that researches the intersectional complexity and instability of social categories.

We have identified several concerns relating to trans, gender diverse, and I/VSC people. The Review's model of immutable, binary, ‘biological’ sex cannot account for variation and change in biological determinants of sex or for factors captured in a more dynamic, sex contextualist approach that recognises sex as bi- or multi-modal. It promotes misgendering and limiting possibilities for trans, gender diverse and I/VSC people's sex or gender to be accurately recorded, presenting this approach as offering clarity and rigour. Finally, we see it as endorsing practices that contravene ethical principles of autonomy, dignity, rights and welfare, particularly for trans, gender diverse and I/VSC people. Taken together, we argue that the Review has the effect of discrediting and refusing trans existence and trans lives and agree with the Lemkin Institute for Genocide Prevention's (2025c, n.p.) assessment that compliance with its recommendations could lead to ‘forced outing of transitioned people’.

SR1 is already having impact internationally. For example, it is cited authoritatively in Trump administration reports, including on ‘Pediatric Gender Dysphoria’, which moves against gender-affirming care for transgender children. There, SR1 bolsters that report's excision of the terms ‘gender’ and ‘gender identity’ and its commitment to biological, ‘sex-based language’ (US Department of Health and Human Services 2025, 39-40). SR1, we argue, must be seen as contributing to reactionary interventions that erase the realities and constrict the socio-spatial worlds of trans adults and children (Gill-Peterson 2018).

Adopting SR1 would also facilitate damaging practices of knowledge-making. To use ‘immutable’ and ‘binary’ sex ‘at birth’ to categorise people constitutes a wilful misrecognition of a core part of most people's understanding of themselves (their sex and gender). Additionally, if this government-commissioned review removes the flexibility of healthcare and other agencies to operationalise the multidimensional category of sex in ways they deem most appropriate, then, we argue, the state's ability to protect its population's health and well-being might be undermined (see Currah 2025). How sex is classified and regulated across different domains profoundly affects patterns of racial, economic and social inequality (Currah 2022).

In response to the Supreme Court judgement and the EHRC's response, 533 geographical scholars called for urgent action to protect the rights of trans geographers and UK trans people (RGS–IBG 2025; Todd 2025). Urgent action also demands pushing back against the Review's far-reaching implications for social science. Social scientists and the wider research community should be particularly concerned by SR2's (16–17) call for universities and funding bodies to reform ethical committees and guidance, amend terms of reference to state they have no remit to review sex conceptualisations and language, and in certain cases remove ethical review requirements from certain research. Ethics committees, we argue, should be more involved than ever in reviewing protocols that adhere to such proposals as those SR1 presents. Government overreach into ethics committees’ operationalisation raises serious questions about academic autonomy and the independence of scientific inquiry. For social scientists to allow the Sullivan Review's recommendations to become woven into ethics, methods, funding protocols and datasets on which many draw would mean complicitly accepting a system that, we argue, excludes and harms trans and gender diverse people, undermines critical scholarship, and contravenes intersectional feminist praxis that many have fought to embed across disciplines.

The Review's calls to order attempt to install a reality wherein trans and gender diverse people are erased; queerness is narrowly defined or effaced; rigid binaries of sex, and a rigid separation of sex and gender are enforced; ethical oversight is regressed; and ethics are transgressed in the name of a particular kind of science authorised as ‘mainstream’. It behoves all social scientists to refuse the Sullivan Review's vision.

Funding

This work was supported by Leverhulme Trust (grant no. ECF2023-481).

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analysed during the current study.