The survey definition

The number in the headline comes from a question, not a diagnosis.

What the question does

The major continuous household surveys — principally the Annual Population Survey and the Family Resources Survey — arrive at a disability count through self-report. Respondents are asked whether they have a physical or mental health condition or illness that has lasted or is expected to last twelve months or more, and whether that condition reduces their ability to carry out day-to-day activities. Both parts must be answered affirmatively for the respondent to be counted as disabled.

That two-part structure is not accidental. It mirrors the shape of the legal test in the Equality Act 2010, which requires both a physical or mental impairment and a substantial adverse effect on normal day-to-day activities. The survey question is not that test, and it carries no legal weight; but the drafters aligned the wording closely enough that the resulting count is treated, in policy documents and parliamentary debate alike, as a reasonable approximation of the Equality Act population.

The definition is therefore self-reported by design. No clinician verifies the condition; no assessor measures functional capacity. The respondent decides whether their circumstances match the question, and the count aggregates those decisions. This is not a flaw to be corrected — it is the mechanism. It means the number reflects how people understand and identify their own situation, which is precisely the kind of information a household survey is built to capture.

The consequence is sensitivity to wording. When the Office for National Statistics revised the question used in the Census, or when survey designers have tested alternative phrasings, measured prevalence has shifted — sometimes substantially — without any underlying change in the population. That the wording decides the number is not a metaphor; it is a documented empirical finding.

The survey definition also differs from the administrative count. A person receiving Personal Independence Payment has been assessed under a separate functional instrument whose criteria do not map directly onto the survey question. The populations overlap but are not identical, and comparing the two figures as though they measure the same thing produces error.

What the survey definition offers — and the reason it remains the dominant source for disability prevalence — is breadth: near-universal coverage of the household population, a consistent instrument across years, and a count that is not contingent on whether someone has applied for, or been awarded, anything.