Food Swaps to Improve the Healthfulness and Reduce the Carbon Footprint of Grocery Purchases: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Food Swaps to Improve the Healthfulness and Reduce the Carbon Footprint of Grocery Purchases (NCT06648226), a randomized na study. The primary outcome is Healthfulness of Product Selection.

2. Clinical Question

The primary measure was the average Ofcom Nutrient Profiling Model score for selected products, with higher scores indicating healthier selections. The stated objective also includes reducing the carbon footprint of grocery purchases, but the registry record does not report a carbon-footprint outcome result among the posted primary results. The question addressed by the detailed primary results is therefore about the healthfulness score, not whether carbon footprint changed.

3. Trial Design

The registry record describes a randomized, parallel-group behavioral trial with no masking. It lists four groups: health swaps, climate swaps, combined health and climate swaps, and control. The study status is completed, and the recorded enrollment is 1201. The start date is 2025-05-07 and the primary completion date is 2025-06-25. The registry record does not report the allocation ratio or the methods used to generate or conceal the random assignment.

Shopping took place in an online grocery-store environment. Health swaps paired nutrition grades with replacement suggestions for products having a better nutrition grade. Climate swaps used climate grades and suggestions for products with a better climate grade. The combined condition displayed both types of grade; its suggestions were for products better than the original choice on at least one dimension and not worse on either. Control participants received no labels or swaps.

The primary outcome represents participants’ selections at the 1-week follow-up (visit 2) and the 2-week follow-up (visit 3). Each participant’s outcome was the mean across those visits when data were available at both; if one visit was missing, the nonmissing visit was used. The analysis population included participants with an outcome at either visit. The arm-level outcome counts differ, so the reported primary summaries do not use the same number of participants in every group.

The registry record mentions survey measures and lists secondary measures related to nutrient density, spending, and thinking about health. The protocol’s original sample-size justification, detailed randomization procedures, and any interim-analysis plan are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Analysis population included participants with nonmissing outcome data at either visit 2 or visit 3.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Healthfulness of Product Selection, measured over 1-week follow-up (visit 2), and 2-week follow-up (visit 3).

The primary endpoint is the healthfulness of participants’ product selections, measured as the average Ofcom Nutrient Profiling Model score of products selected in the shopping task. The scale ranges from 0 to 100, and a higher score represents a healthier selection. The outcome is expressed as a mean and standard deviation for each group.

The endpoint combines information across the two follow-up shopping visits. For participants with data at both visit 2 and visit 3, the outcome is their mean across those visits. If data were missing at one of the visits, the value from the nonmissing visit was used. Participants with no outcome at either follow-up were outside the stated analysis population.

This measure directly reflects the study’s primary healthfulness question: whether the labels and swap suggestions were associated with the nutritional profile of selected products. It is a score for selections in a structured online task, not a measure of overall diet, health status, or long-term behavior. Although the study objective also refers to the carbon footprint of purchases, the registry record does not report a carbon-footprint result alongside the primary endpoint. The endpoint therefore answers a narrower question than the full set of stated study aims.

7. Results Posted in the Registry

• Health Swaps: mean 64.00 versus 57.90 on control; difference 6.10 (95% CI 4.94 to 7.26), standardized effect d = 0.88; z-test p = <0.001.
• Climate Swaps: mean 60.10 versus 57.90 on control; difference 2.20 (95% CI 1.14 to 3.26), standardized effect d = 0.34; z-test p = <0.001.
• Combined Health and Climate Swaps: mean 62.60 versus 57.90 on control; difference 4.70 (95% CI 3.46 to 5.94), standardized effect d = 0.64; z-test p = <0.001.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Calorie Density of Participants' Grocery Selections

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Health Swaps: mean 166.80 versus 200.50 on control; difference -33.70 (95% CI -45.05 to -22.35), standardized effect d = -0.49; z-test p = <0.001.
  • Climate Swaps: mean 182.20 versus 200.50 on control; difference -18.30 (95% CI -29.37 to -7.23), standardized effect d = -0.27; z-test p = 0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Health Swaps than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Climate Swaps than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Sugar Density of Participants' Grocery Selections

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Health Swaps: mean 6.40 versus 8.00 on control; difference -1.60 (95% CI -2.56 to -0.64), standardized effect d = -0.28; z-test p = 0.001.
  • Climate Swaps: mean 6.90 versus 8.00 on control; difference -1.10 (95% CI -2.02 to -0.18), standardized effect d = -0.20; z-test p = 0.019.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Health Swaps than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Climate Swaps than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Sodium Density of Participants' Grocery Selections

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Health Swaps: mean 264.10 versus 317.30 on control; difference -53.20 (95% CI -81.30 to -25.10), standardized effect d = -0.31; z-test p = <0.001.
  • Climate Swaps: mean 314.40 versus 317.30 on control; difference -2.90 (95% CI -32.96 to 27.16), standardized effect d = -0.02; z-test p = 0.850.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Health Swaps than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Climate Swaps and the control. The interval gives the range of effects that remain compatible with the data.

The registry record lists secondary measures covering calorie, sugar, sodium, saturated fat, fiber, and protein density, spending, and thinking about health. The detailed arm summaries include calorie, sugar, and sodium density. Across those measures, the pattern is not uniform across all swap groups and outcomes: lower average nutrient-density values appear in several comparisons, while the degree of alignment with the primary healthfulness pattern varies by measure and intervention.

These secondary measures help describe what the healthfulness score may reflect, because nutrient density and a composite nutrition score address related but nonidentical features of selected products. They do not substitute for the primary endpoint, and they do not establish that every nutritional dimension changed in the same way. Spending and thinking about health are also distinct outcomes rather than direct measures of the primary score.

The registry record identifies 30 outcome measures, while the reported analysis listing contains no analyses. The available summaries should therefore be treated as exploratory context unless a prespecified analysis plan says otherwise. When many outcomes and comparisons are considered, some apparently notable patterns can occur by chance. Unadjusted secondary results carry less weight because the registry record does not report a multiplicity adjustment or a hierarchy defining which secondary findings were confirmatory.

The study’s stated objective includes carbon footprint, but the secondary measures named in the registry record and the detailed outcome results do not provide a reported carbon-footprint estimate. The secondary results therefore add context about nutritional selection characteristics, but they cannot establish the environmental effect of the swaps. Nor can results from an online task alone show that the same selection patterns occurred in real-world purchases or persisted over time.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Calorie Density of Participants' Grocery Selections
• Sugar Density of Participants' Grocery Selections
• Sodium Density of Participants' Grocery Selections
• Saturated Fat Density of Participants' Grocery Selections
• Fiber Density of Participants' Grocery Selections
• Protein Density of Participants' Grocery Selections
• Total Spending on Participants' Grocery Selections
• Thinking About Health

9. Statistical Interpretation

10. Statistical Methods Explained

The primary summaries are means and standard deviations because the endpoint is a continuous score. A group mean describes the average healthfulness score among participants in that group. A standard deviation describes how dispersed individual participant outcomes are around that group mean. It is not the same as the standard error, which describes the precision of an estimated mean or contrast.

The comparisons are expressed as mean differences between each swap group and control. A positive difference means the swap group’s average score is higher than control’s, in keeping with the scale definition that higher scores indicate healthier selections. The reported contrasts are unadjusted recalculations from the posted summary statistics. They are not evidence that a covariate-adjusted model or another sponsor analysis produced the same estimates. The registry record does not report a detailed model specification for the primary analysis.

The p-values attached to the comparisons are also unadjusted. In general, a p-value summarizes how surprising data at least as incompatible with a null hypothesis would be if that null and the calculation’s assumptions were correct. It is not an effect size and does not describe the probability that an intervention is beneficial. Because the registry record does not report a multiplicity strategy, the unadjusted p-values should not be read as accounting for the several group comparisons and outcome measures.

The endpoint combines observations across follow-up visits at the participant level. When both visit values were available, the participant’s mean across visits was used; when only one was available, that nonmissing value was used. This rule makes the endpoint usable for participants with a value at either follow-up. It does not explain why values were missing, establish that missingness was unrelated to outcomes, or replace a missing-data sensitivity analysis. Such analyses are not reported in the registry record.

The analysis population is participants with nonmissing primary outcome data at visit 2 or visit 3. This is distinct from the total enrollment figure. The registry record gives arm-specific outcome counts, but does not report a detailed account of participants without usable outcome data, their reasons for missingness, or how missingness differed across groups. Those details matter because missing outcomes can affect how representative the analyzed group is of everyone randomized.

Confidence intervals give a range of values compatible with the observed estimate under the assumptions of the calculation. Here, the intervals accompanying the mean differences quantify uncertainty around each unadjusted contrast. Their width reflects both the variability of participant scores and the amount of information available. The intervals do not capture every possible source of uncertainty, including choices about the analysis model, missing data, or multiple comparisons.

Over-dispersion is a concern when outcome variation exceeds what a statistical model assumes. In general, failing to account for over-dispersion can make standard errors too small and confidence intervals too narrow. The primary endpoint is treated as a continuous measure with reported means and standard deviations; the registry record does not describe a count model or report an over-dispersion assessment. This concept is therefore general context, not a claim about a diagnosed feature of this analysis.

The registry record lists 30 outcome measures and the posted record reports no analyses in its analysis listing. It does provide arm-level outcome summaries and the comparison values. The absence of a listed analysis does not reveal which analyses were planned, whether a formal statistical analysis plan existed, or whether interim analyses were conducted. Those details are not reported in the registry record.

The sample-size illustrations use a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. They examine the stated fractions of the observed effect. These are planning calculations based on observed effects, not a reconstruction of the trial’s original design assumptions. The original sample-size justification is not reported in the registry record.

11. Confidence Intervals and Precision

The 95% confidence intervals for the unadjusted primary mean differences are 4.944 to 7.256 for health swaps versus control, 1.142 to 3.258 for climate swaps versus control, and 3.463 to 5.937 for combined swaps versus control. Each interval is on the healthfulness score scale. The interval for a contrast represents uncertainty about that group difference, not the range containing most individual participants’ scores.

All three intervals lie on the positive side of zero, and the posted unadjusted p-values are < 0.001. Under the assumptions behind these recalculations, the estimates are more precise than a result with a much wider interval would be. Precision is not the same as importance: the interval does not decide whether a difference is meaningful for consumers, nor does it establish that the result persists outside the online task.

The primary outcome standard deviations range from 6.3 to 8.1 across the four groups. These values describe participant-to-participant variation, whereas the confidence intervals describe uncertainty in the group contrasts. Keeping those interpretations separate avoids treating the spread of individual scores as if it were the precision of the estimated mean difference.

The intervals are unadjusted and do not reflect a reported correction for the several comparisons and outcomes. The registry record does not report a prespecified multiplicity procedure, sensitivity intervals under alternative missing-data assumptions, or an adjusted model-based estimate. Thus, the stated precision is conditional on the simple unadjusted comparison and the available outcome data. It should not be interpreted as incorporating every uncertainty arising from analysis choices or incomplete follow-up.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size illustrations give participants per arm for 80% power and participants per arm for 90% power. They are not total event counts: the outcome is continuous, and the calculations describe the number of participants in each of two equally allocated groups. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each row considers an effect equal to the stated fraction of the observed effect.

The 80% power column describes a planning scenario in which a study with the listed participants per arm would have an 80% chance of detecting the assumed effect at the stated significance level, if the assumptions used in the calculation hold. The 90% column uses a higher target chance of detection and therefore generally calls for more participants. These percentages describe hypothetical design performance, not the probability that the observed trial result is correct.

As the assumed effect becomes smaller, more participants are needed to distinguish it from random variation under the same planning assumptions. The rows that use a fraction of the observed effect illustrate that relationship. Very large requirements can arise when the observed contrast is small relative to the variability in the outcome; they are consequences of the calculation, not evidence that such a study was conducted.

This is not observed power. The illustrations are planning calculations based on the observed effects, not a measure of the probability that this completed trial found a real effect. The trial’s original sample-size justification and original design assumptions are not reported in the registry record.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 6.100; pooled standard deviation σ = 6.965.
2. n = 2 × (1.96 + 0.842)² × 6.965² / 6.100² = 20.5, rounded up to 21 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- 6.1                      # observed difference in means
sigma <- 6.96455310841981                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(6648226)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 21 (80% power), 28 (90% power)
 75% of observed effect: n per arm = 37 (80% power), 49 (90% power)
 50% of observed effect: n per arm = 82 (80% power), 110 (90% power)
power.t.test: 21.5 
Simulated power at n = 21 per arm: 0.787 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Multiplicity, Missing Data and Analysis Populations

The registry record identifies 30 outcome measures and three swap-group comparisons against control for the primary endpoint. It does not report a multiplicity adjustment, a hierarchy of primary and secondary hypotheses beyond the named primary outcome, or a statistical analysis plan specifying how multiple comparisons were to be handled. The listed p-values are unadjusted. With multiple outcomes and comparisons, interpreting each result as if it were the sole planned test can overstate the strength of evidence.

There are two follow-up visits contributing to the primary outcome. The stated rule averages the participant’s values across visits when both are available and uses the single nonmissing visit when one is missing. Participants without an outcome at either follow-up are not part of the defined primary analysis population. The registry record does not report the number or reasons for missing outcomes, a comparison of missingness across groups, or sensitivity analyses addressing missing data.

The registry record describes a randomized trial, but it does not report whether the primary comparison followed an intention-to-treat principle, how deviations from assigned conditions were handled, or whether exclusions were made after randomization. It also does not report a detailed analysis model, covariate adjustment, interim-analysis plan, or other prespecified analysis decisions. These missing details limit assessment of whether the reported summaries and recalculated comparisons match the intended analysis.

The completed status and enrollment figure provide context, but enrollment alone does not establish the number included in each outcome analysis. The posted arm counts apply to participants with primary outcome data and are not a full accounting of follow-up or missingness. The registry record does not report the trial’s original sample-size justification. Taken together, these gaps mean the available results can be described, but the analysis cannot be fully audited against a reported prespecified plan.

15. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

For visits 2 and 3, the registry reports zero serious adverse events, zero other adverse events, and zero deaths in each of the health-swap, climate-swap, combined-swap, and control groups. The reported numbers at risk for these groups are 300, 301, 300, and 300, respectively. For the no-intervention condition at visit 1, the record reports zero serious adverse events, zero other adverse events, and zero deaths among 1201 participants at risk.

These figures describe the events recorded in the registry for the specified groups and periods. A zero count means that no events of those categories were reported in those records; it does not by itself demonstrate that no event could occur, or establish a comparative safety profile. The registry record does not report serious-event terms, and no event detail is available to characterize possible causes or severity beyond the stated categories.

The figures should also be read in context: the interventions were behavioral features in an online grocery-shopping task, and the primary outcome concerns selected products. Safety counts do not answer the study’s questions about nutritional selection or carbon footprint.

16. What the Result Does, and Does Not, Mean

The posted primary results show higher average healthfulness scores in the three swap groups than in control during the online grocery-shopping task. The unadjusted recalculated mean differences are positive for health swaps, climate swaps, and combined swaps, and their confidence intervals do not include zero. These estimates describe the participants and outcome data included in the registry’s primary analysis.

The result does not establish that any particular intervention caused lasting changes in food purchasing or eating. The endpoint is based on product selections in an online task, not documented long-term consumption or clinical health outcomes. The registry record does not report whether participants later purchased the selected products or whether their behavior persisted after the follow-up period.

The result also does not answer the full environmental question in the study objective. The registry record does not report a carbon-footprint result among the posted primary results. A higher nutrition-profile score cannot be used as a substitute for an environmental-impact measurement, and the detailed secondary results do not provide that missing outcome.

Interpretation is limited by unadjusted comparisons, the absence of a reported multiplicity strategy, and missing details about the analysis plan and missing outcomes. The confidence intervals communicate uncertainty around the estimated score differences under the recalculation assumptions; they do not resolve these design and reporting uncertainties. The registry record also does not report a clinical threshold for a meaningful change on the score scale.

Accordingly, the data support a description of higher average scores among participants assigned to swap conditions than among control participants in this study setting. They do not establish the size or durability of any effect in routine shopping, the consequences for health, or changes in carbon footprint.

17. Why This Trial Matters Statistically

This trial is statistically interesting because it tests distinct ways of presenting information and suggesting substitutions within the same online shopping environment. Comparing health-focused, climate-focused, and combined approaches with a control condition allows the reported group summaries to describe whether the primary score differed across these strategies. The combined condition also has a specific rule for substitutions: a product should improve at least one dimension without worsening the other.

The primary outcome is a continuous score, so group means and mean differences make the direction and scale of the observed contrasts interpretable. Confidence intervals add information about precision that a p-value alone cannot supply. The arm-level standard deviations also show that individual scores vary within each group, a relevant context when interpreting average differences.

The trial highlights why endpoint choice matters. The stated goal includes both healthfulness and carbon footprint, but the posted primary result is a nutrition-profile score. Without a reported carbon-footprint estimate, conclusions about environmental impact remain outside the available outcome evidence. Similarly, an online selection task is a specific behavioral setting and should not be conflated with actual purchases or long-term diet.

Finally, the registry record illustrates the value of transparent reporting of planned analyses, missing-data handling, and multiplicity decisions. The means and unadjusted comparisons are informative descriptions, but the absence of those details limits how confidently they can be interpreted as confirmatory evidence. Statistical interpretation depends not only on estimates, but also on how outcomes and analyses were selected and reported.

18. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The study evaluates selections in an online grocery-store environment. The registry record does not report whether those selections became actual purchases, how long any behavioral changes lasted, or whether participants’ usual shopping conditions were comparable to the task. Generalization from the task to routine purchasing is therefore uncertain.

The primary endpoint measures nutrition-profile scores, while the stated study objective also concerns carbon footprint. A carbon-footprint result is not reported in the registry record, so the environmental part of the objective cannot be assessed from these outcomes. The healthfulness score is not an environmental measure.

Outcome values combine data from two follow-up visits using the nonmissing value if one visit is unavailable. The registry record does not report the extent, reasons, or group distribution of missing outcomes, nor a sensitivity analysis. The arm-specific analysis counts also indicate that the outcome summaries are based on subsets defined by available data rather than simply the enrollment total.

Comparisons are unadjusted recalculations, and the registry record does not report a multiplicity strategy or detailed statistical analysis plan. It also does not report the original sample-size justification, allocation ratio, or interim-analysis plan. These reporting gaps limit assessment of how the posted comparisons relate to the planned confirmatory analysis.

The record mentions survey measures and a broad set of secondary outcomes, but the detailed results include only some of those measures. The registry record does not provide reported analyses for the listed secondary results. This makes it difficult to evaluate all intended dimensions of the study or to distinguish planned confirmatory results from exploratory summaries.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Food Swaps to Improve the Healthfulness and Reduce the Carbon Footprint of Grocery Purchases — NCT06648226.
• Registry record last updated 2026-06-24; status last verified 2026-05.
• Results first posted 2026-06-01.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

22. Record Summary

The registry record describes a completed randomized, parallel-group behavioral trial with no masking and enrollment of 1201. Participants were assigned to health swaps, climate swaps, combined health and climate swaps, or control in an online grocery-shopping environment.

The primary endpoint was the mean Ofcom Nutrient Profiling Model score for selected products across the 1-week and 2-week follow-ups, with higher scores indicating healthier selections. The posted arm means were higher in each swap group than in control. Unadjusted recalculated mean differences versus control were 6.1, 2.2, and 4.7 for health, climate, and combined swaps, respectively.

The registry record does not report a carbon-footprint result, a multiplicity strategy, or the original sample-size justification. The primary summaries describe healthfulness scores in an online task and do not establish effects on lasting purchasing, diet, clinical outcomes, or environmental impact.