The interesting quantity here is the distance between the three numbers rather than any one of them. Moving from the socially comparative description to the minimum-standards description lifts agreement by 27 points, 41% to 68% [4][5][16]. Moving on to the case where a person cannot eat and live without borrowing adds only 12 more [6][16]. Consensus is cheap at the subsistence end and scarce at the relative end, which is where Syrda reports the ideological spread at its widest [8].
Put that spread on the sample. The gap between mildest and severest description is 39 percentage points [14], which across 2,642 respondents is roughly 1,030 people [15] who put borrowing-to-eat inside the boundary and exclusion from ordinary living standards outside it. That arithmetic assumes answers nest, that nobody called the mild case poverty while refusing the severe one, and the release does not report the crosstab, so read it as an upper bound on the definitionally divided group.
Notice what the three prompts actually are: sentences about circumstances, with no incomes, household sizes or regions attached [3]. That is the right design for this question, because fixing a pound figure would import the very argument the study wants to observe. It is also the design's limit. Two respondents reading "having enough to eat and live but not enough for other necessities" [3] may be picturing quite different budgets, and the survey cannot separate a narrow boundary from a generous mental image of the same sentence.
The paper's subtitle claims definitional mediation [1], and the data are a single 2022 wave [2]. Boundaries and beliefs about the scale of poverty do move together in it [11], and Syrda's reading is that the disagreement begins with the definition and flows outward to the estimate [12]. One cross-section cannot order those two. Someone already convinced that poverty in Britain is extensive has every reason to widen the label as well, and the correlation looks identical either way. The arrow is a modelling decision, and it is worth saying so out loud.
What travels better is the party result: judgments tracked left-right position and support for welfare independently of which party the respondent voted for [10]. Vote choice is the variable most easily to hand in a newsroom or a focus group, and on this evidence it is the wrong one.
The thing this does not tell you is whether anyone's line moves. Nothing in the survey pushes a respondent toward a different boundary, so the study measures where lines sit, not how firmly they are held. Syrda asks researchers, policymakers and journalists to state which of the three definitions is in play [13], on the grounds that explicit assumptions clarify what is being argued about even when they do not resolve it [13]. On these numbers, a reader handed the word "poverty" with no definition attached is being handed a figure that somewhere between 41% and 80% of the public would recognise as poverty at all [4][6].
