The UK has spent the better part of a decade redesigning its supermarkets from the regulator’s office. The 2022 Food (Promotions and Placement) rules already dictate where stores may put crisps and chocolate, banning the worst offenders from checkouts, aisle ends, and entrance zones. Now a research team funded by the National Institute for Health and Care Research is asking ministers to take the next step: an explicit rule, for every large food store over 2,000 square feet in England, on where the carrots have to live.
The evidence base for that ask, published this month in PLOS Medicine, is a 36-store non-randomized trial in a single discount chain whose primary household-purchasing signal never reached significance.
The WRAPPED trial, led by Christina Vogel at City St George’s, University of London and Janis Baird at the University of Southampton, was structured cleanly enough. Eighteen stores got the intervention: the fresh produce section moved to the front and the range expanded. Eighteen kept their existing layout. The team tracked store sales, the loyalty-card baskets of roughly 580 enrolled households, what the women in those households reported eating, and how often produce got thrown out at home.
The figure the university press release and the follow-on coverage have led with is the till receipt. At the moment the layout changed, intervention stores were ringing up about 2,525 extra portions of fruit and vegetables per store, per week, against the model-predicted counterfactual (95% CI 775 to 4,115). In standardised units that was 0.32 SDs at zero, a clean and statistically significant bump (p = 0.002). By three months the effect had fallen to 0.23 SDs and lost significance (p = 0.10). By six months it was 0.18 SDs (p = 0.29). The retail signal the policy ask rests on did not survive a quarter of the year.
The household question is where the case has to land, because telling Tesco where to put the carrots only matters if families take more carrots home. The primary household-purchasing outcome was a dichotomized loyalty-card variable: in any given week, did the household buy any fresh produce at all? Intervention households were 0.6 percentage points more likely to do so at three months (p = 0.83) and 3.3 percentage points more likely at six months (p = 0.23). Both confidence intervals straddled zero comfortably. The women in the cohort registered a 0.25 SD gain on a 20-item food-frequency questionnaire at six months (95% CI 0.10 to 0.40), a modest improvement on a self-report instrument. And the vegetable-waste signal, measured by telephone survey asking how often respondents threw vegetables away because of spoilage or over-preparation, rose 0.20 events per week at six months (95% CI 0.07 to 0.32). More produce crossed the checkout. More vegetables, by household report, went in the bin.
This is the shape these interventions almost always take. A clean bump at the till that decays. A diet-quality nudge on a self-report instrument. A purchasing signal at the household level that does not separate from noise. The WRAPPED trial does not show the redesign failed. It shows it worked the way choice-architecture interventions work: a little, for a while, with leakage at every stage between the shelf and the dinner plate.
What deserves a hard pause is the gap between that finding and the policy ask delivered in the same paper. The authors recommend that the Food (Promotions and Placement) regulations be expanded to “require the positioning of fresh produce sections near store entrances in all large food stores” over 2,000 square feet in England. A six-month look at 36 stores of one unnamed discount chain, with a primary till effect that lost statistical significance by week thirteen and a household-purchasing endpoint that never reached it, is being offered as the evidentiary basis for a national mandate on how every large supermarket in the country arranges its floor space.
The funding flow is worth naming plainly, because it is the cleanest version of the dynamic. The trial was paid for by the NIHR Public Health Research Programme and the NIHR Southampton Biomedical Research Centre. Both are UK government bodies whose explicit remit is generating evidence for public-health policy. The output is a paper that delivers, alongside its results, a specific regulatory ask back to the same government. That is the cycle: state-funded evidence underwriting state-issued layout rules, with the supermarket chain involved kept anonymous. The supermarket put up no money. Two co-authors carry declared interests the team deemed unrelated to the work. Janet Cade directs Dietary Assessment Ltd, a company that develops dietary-assessment tools of the same general family the study relied on. Cyrus Cooper holds consulting relationships with pharmaceutical firms. Readers can weigh those disclosures themselves.
Two features of the trial design carry weight. First, this was not a randomised trial. The paper says so. Stores were drawn from the chain’s list of locations already scheduled for refurbishment and structural changes, then matched to controls. That is a pragmatic choice in a commercial setting, but it cannot rule out the possibility that refurbishment-scheduled stores were trending differently from the others before a single carrot moved. Second, the study window ran from March 2018 to May 2022. That is the COVID shock, the Brexit-driven supply disruption, and the front edge of the cost-of-living squeeze, all stacked inside the analysis. UK household fruit purchasing fell 7.2% and vegetable purchasing fell 5.3% at the population level across that period. The authors lean on that turbulence to explain why their till effect faded; the same turbulence is why any household signal would have been hard to pull out of the noise even if one had been there to find.
The question worth pressing is what kind of result would have to land before publicly funded nudge research stopped reaching, in the same breath, for the regulatory lever. A 0.32 SD till bump that loses significance inside a quarter, a household-purchasing endpoint that never reaches significance at all, and rising vegetable waste at home is not a null result. It is also not the kind of finding that, in a saner regulatory culture, would justify telling every large food store in England how to lay out its floor. Watch what Westminster does with it. The architecture for the next ratchet is already in place, and the paper recommending it is already on the minister’s desk.