Guide · 16 July 2026

Do you need a single-sex service? The eight-gate assessment explained

Is this a single-sex service, does it need to be, what options are there, and how do we operate and maintain it — counting the data-protection cost of policing gender and presentation? The eight-gate assessment forces those questions in the right order.

By Joanne Lockwood · 12 min read

Four questions sit in front of any manager considering single-sex provision: is this service single-sex? Does it need to be? What options are there? And how would we operate and maintain it — including the data-protection cost of policing gender and presentation? The eight-gate assessment is the discipline that forces those questions in order, so single-sex is reached only when it is genuinely needed and its costs are understood — never assumed.

This article is an explainer of the gate sequence, not a substitute for the full assessment. The topic hub on lawful trans inclusion in single-sex services sets out the framework at a glance; the sibling single-sex spaces manager decision framework gives a manager the four-step process for the day. What follows is the structured rigour underneath both. Take specific legal advice before finalising any policy built on this assessment.

Is it, and does it need to be?

Start by establishing what the service actually is. Many services described as “single-sex” are, in law, mixed-sex services with a women-centred focus — and that is often the better design. The EHRC’s 2026 statutory Code of Practice for services — in force from 5 August 2026 — frames a mixed, open-to-all service as the ordinary starting point, with single-sex provision a justified exception, never an assumption. The burden of justifying a restriction sits with the provider, not with the person seeking access. Schedule 3, Part 7 of the Equality Act 2010 sets out the exceptions that permit a departure; they are permissive, never mandatory. Nothing in the Supreme Court’s 2025 holding in For Women Scotland — that “sex” in the Act means biological sex — creates a duty to exclude. The judgment clarified the definition; it did not change the default.

The options are a spectrum, not a binary, and most “should this be single-sex?” questions are better answered further up the list than the manager first assumes:

  • Mixed, open to all — the lawful default.
  • Separate-sex — both available, side by side.
  • Single-occupancy or third-space — individual, lockable provision anyone can use.
  • Eligibility by need, risk or purpose — a specialist service whose criteria are about vulnerability or purpose, not sex.
  • Single-sex on the Code’s terms — a restriction relying on a Schedule 3 exception, justified by proportionality.

Getting that backwards — treating the service as single-sex by default and requiring the individual to earn their way in — is the single most consequential error a decision-maker can make here.

The cost of going single-sex: policing gender and presentation

This is the question most often skipped, and it belongs upfront. A service that is single-sex by policy has to be operated as single-sex at the door — and that means deciding who fits, which in practice means judging appearance and presentation, or demanding evidence of sex. No UK document reliably evidences biological sex for service access. Appearance is unreliable. Challenging a person at an ordinary service door is itself a dignity risk and, the moment staff start judging or recording a person’s sex or gender presentation, a data-protection risk: sex and gender-reassignment data are special-category data, and the ICO’s guidance on inference of special-category data is engaged the moment those judgements are made or recorded.

So the operational cost of single-sex is a liability the provider takes on: a door that must be policed, a judgement that cannot be made reliably, and a processing risk that arises from the policing itself. That cost belongs in the “does it need to be?” decision, weighed before the restriction is chosen — not discovered after, when a complaint or a subject-access request arrives. A restriction that cannot be operated lawfully at the door cannot be lawfully maintained, however well the proportionality paperwork reads.

The eight gates

Each gate is a question, and each is a pass or a flag. A flag does not halt the assessment — it is recorded, and its consequence is carried forward into the proportionality reasoning. What the sequence prevents is the quiet drift to a restriction without the earlier gates ever being engaged honestly.

Gate 1 — Is single-sex genuinely needed?

Before reaching for a restriction, ask whether the need is real or assumed, and whether a mixed, separate-sex, single-occupancy, or eligibility-based model would meet the purpose. Mixed provision is the lawful default; single-sex is a departure that must be justified.

  • A pass looks like a clear service rationale with a sex-based restriction genuinely engaged — not a default reached because it feels safer, is easier, or is being demanded by pressure.
  • A flag looks like the assessment having moved toward restriction without establishing why the service needs to be single-sex at all. An exclusion recommendation is structurally unreachable from a flagged Gate 1.

Gate 2 — Is there a lawful basis?

The Equality Act’s single-sex exceptions set out the conditions under which a service can lawfully depart from mixed provision — including where only one sex needs the service, where joint provision would be insufficiently effective or impracticable, in a hospital or special-care setting, where a reasonable objection to the opposite sex’s presence arises (including in undress or vulnerability), and where physical contact is involved.

  • A pass looks like at least one condition identified and named, in plain terms tied to the service in front of you.
  • A flag looks like no condition genuinely applying — which means the single-sex framing is not available in law and the service is, in law, mixed. The right response is to design it as mixed, not to reach for a condition that does not fit.

Gate 3 — Is the restriction proportionate?

This is the core. A single-sex restriction has to be a proportionate means of achieving a legitimate aim — the four-limb test from Bank Mellat v HM Treasury (No. 2) [2013] UKSC 39: a legitimate aim, a rational connection, a minimal-impairment assessment of less intrusive alternatives, and a balance where the benefit outweighs the impact on the affected group. The impact on trans people sits inside that balance — mandatory, not optional.

  • A pass looks like proportionality reasoning recorded — ideally a proportionality file (Gate 8) — with the impact on the affected group weighed inside it and the less-intrusive alternatives genuinely considered.
  • A flag looks like “the Code says biological sex, therefore exclude” with no proportionality run; or the balance flattened to a single side. Without this gate, the single-sex justification fails as a matter of law.

Gate 4 — Does anyone end up with nowhere to go?

A routing rule that leaves a person with no usable provision — excluded from one route by others’ reasonable objection and from another by their own presentation — fails proportionality as designed. The Code’s dignity floor is clear: for a necessary service, it is very unlikely to be proportionate to leave a person with no facility they can use at all.

  • A pass looks like a usable, dignified alternative existing for anyone a routing rule would otherwise leave stranded — an individual room, a lockable facility, additional provision, mixed access to a comparable service.
  • A flag looks like a rule that can produce a no-usable-provision outcome. The remedy is an alternative, not a tighter rule.

Gate 5 — How are objections handled?

“Reasonable objection” is not “anyone feels uncomfortable” or “someone complains”. Genuine privacy, safety and dignity concerns must be distinguished from prejudice, ideological objection, generalised anxiety, or mere conduct-free presence. The objection has to be assessed against the service context — not treated as automatic.

  • A pass looks like the policy requiring objections to be assessed, not triggered — with the distinction between a genuine dignity concern and a heckler’s-veto drawn explicitly.
  • A flag looks like the policy delegating access to whoever objects loudest — a rule that hands the most hostile or anxious user a veto over everyone else.

Gate 6 — How is it operated day-to-day, and what does policing cost?

This is where the cost flagged above is tested against an operating model. Lawful single-sex provision is secured by clear communication — signage, admission and induction — not by verification at the point of access. No UK document reliably proves sex; appearance is unreliable; and it is unlikely to be appropriate to challenge people in ordinary settings.

  • A pass looks like an operating model built on communication and a behaviour-based response — staff respond to conduct, not appearance, and escalate only on clear evidence of an issue. No special-category processing is triggered by routine operation.
  • A flag looks like a policy that contemplates routine challenge, appearance-based suspicion, or document checks — carrying their own legal and data-protection risks without securing the lawful character of the service. A restriction that needs policing the provider cannot lawfully do is a restriction the provider cannot maintain.

Gate 7 — Is the service described honestly?

If a provider relies on the single-sex exception, the service is single-sex on the Code’s terms — and describing it as “women-only, inclusive of trans women” when the exception is being relied on carries risk, because in law the service is then mixed-sex. The wording has to match the legal character.

  • A pass looks like a description matching what the service is in law — either single-sex on the Code’s terms, or honestly framed as a women-centred mixed-sex service, or a specialist service with eligibility based on need, risk and purpose.
  • A flag looks like wording that claims a single-sex character the service does not have in law. The remedy is honest framing, anchored in statute and plain-English principle — not in case names embedded in policy text.

Gate 8 — Is the reasoning documented?

A written policy is usually necessary, and decisions should be recorded with supporting evidence. The proportionality file is the defensibility itself: the rationale, the evidence, the less-intrusive options considered, the impact on the affected group, the alternative access provided, how staff operate it, and the review point.

  • A pass looks like a file that exists before the decision is communicated — written down, dated, showing what was considered and rejected, not only what was decided.
  • A flag looks like signage or policy wording changed without a documented assessment. An undocumented decision is vulnerable on review, however well-reasoned it was.

The no-blanket-rules discipline

Running through all eight gates is a discipline the Code is consistent about: blanket rules are not endorsed. A single-sex restriction is a case-by-case judgement, not a category-level answer, and the proportionality test is inescapably fact-specific. The recent Employment Tribunal decision in Hutchison & Others v County Durham and Darlington NHS Foundation Trust illustrates the point: an employer’s blanket changing-room policy, imposed without consultation or alternatives, was found to constitute harassment and indirect sex discrimination. The failure was not the presence of a trans colleague; it was a single arrangement with no assessment and no alternative — precisely the outcome this discipline exists to prevent.

A worked example (generic)

Consider a generic service — a counselling group for women — where the question arises whether to restrict access on the basis of sex. First, is it single-sex? Many counselling groups are women-centred mixed services already, and that may be the honest answer. If the question is does it need to be, Gate 1 asks whether a sex-based restriction is genuinely needed, or whether an eligibility-by-need model (open to anyone affected by the presenting issue) would meet the purpose. Gate 2 asks whether a Schedule 3 condition applies — most likely the reasonable-objection condition, given the vulnerability inherent in a counselling setting. Gate 3 runs the proportionality test: the legitimate aim, the less-intrusive alternatives (an individual session, a separate group, mixed provision with eligibility criteria), and the impact on the trans women who would lose access. Gates 4 and 5 check whether any routing rule leaves a trans woman with no counselling provision she can use, and filter any objection through the service context. Gate 6 asks how the door would actually be operated — and whether the policing it would require is itself lawful and maintainable. Gates 7 and 8 confirm the service is described honestly and documented in a proportionality file before the decision is communicated. The gates do not tell the provider what to decide; they ensure the decision is reached in the right order, with the right questions genuinely asked.

How the eight gates connect to an EqIA/DPIA

The eight-gate assessment does not stand alone. A single-sex decision is an equality matter that belongs inside an Equality Impact Assessment, and — because operating a sex restriction means processing information about sex and gender presentation — it also belongs inside a Data Protection Impact Assessment. Gates 3 and 4 generate the evidence an EqIA needs: the proportionality reasoning and the impact on affected groups. Gate 6’s operating-model test is where the DPIA’s processing questions start: what special-category data does policing the door create, and is that processing lawful and necessary? Gate 8’s proportionality file is, in substance, the record an EqIA and a DPIA both expect to see. The EqIA every trans inclusion policy needs and the DPIA for trans inclusion set out how those assessments are built; all three should be run together, not bolted on afterwards.

Take this further

The eight-gate sequence is the consulting canon’s structured discipline, and a public explainer can only go so far. For a live decision with real exposure, the full Single-Sex Services Assessment Companion runs all eight gates against your specific facts and produces the documented proportionality file — specialist consulting work, not a self-service wizard. To test a contemplated restriction’s proportionality before you rely on it, the Proportionality Check runs the four-limb test and surfaces the risk level. To build the impact assessment the decision sits inside — including the data-protection assessment for any door-policing — the EqIA/DPIA Wizard takes you through it. For any decision with serious consequences, take specific legal advice — the gates are a discipline, not a substitute for a lawyer.

Take this further

  • Proportionality Check

    Runs a contemplated single-sex restriction through the four-limb proportionality test and surfaces the risk level before you rely on it.

  • EqIA/DPIA Wizard

    Builds the equality and data-protection impact assessment a single-sex decision has to sit inside — including the processing that policing the door would trigger.

  • Consulting

    For a decision with real exposure, the full Single-Sex Services Assessment Companion runs all eight gates against your facts and produces the documented proportionality file.

Sources