https://orcid.org/0000-0002-8705-038X
Figures
Abstract
Background
Modeling studies suggest that simple dietary “swaps” in which just one food is replaced with another could yield both health and environmental benefits, but it remains unclear whether consumers make such changes when prompted. We tested whether recommending swaps improved the healthfulness and reduced the carbon footprint of simulated online grocery selections.
Methods and findings
In May–June 2025, we recruited 1,201 US adults ages 18+ for a randomized parallel trial. Participants completed 3 weekly study visits. In the first visit, participants shopped in a naturalistic online grocery store with no intervention. They were then randomized in parallel to 1 of 4 trial arms: (1) climate swaps (n = 301), (2) health swaps (n = 300), (3) climate + health swaps (n = 300), or (4) control (n = 300). In the second and third visits, participants completed the shopping task in their randomly assigned trial arm. Participants were not masked to their trial arm. Using intent-to-treat analyses, we assessed 2 primary endpoints: the healthfulness (operationalized as the average Ofcom Nutrient Profiling Model score across the selected items [range: 1–100]) and carbon footprint (operationalized as the average kg of carbon dioxide equivalents [CO2-eq] per kg associated with producing the selected items) of participants’ grocery selections. We examined whether outcomes changed from the first (no-intervention) visit to the second and third (intervention) visits in the swaps arms compared to the control arm.
The climate (B = 2.4 on the 1–100 scale, 95% CI [1.1, 3.7]), health (B = 4.7, 95% CI [3.3, 6.0]), and climate + health (B = 4.0, 95% CI [2.7, 5.4]) swaps each improved the healthfulness of participants’ selections compared to control (all ps < 0.001). The health and climate + health swaps both increased healthfulness more than the climate swaps (ps ≤ 0.02) but did not differ from one another (p = 0.34). The climate (B = −3.3 kg CO2-eq per kg, 95% CI [−4.6, −2.1]; p < 0.001) and climate + health swaps (B = −2.8, 95% CI [−4.1, −1.6]; p < 0.001), but not the health swaps (B = −0.4, 95% CI [−1.7, 0.8]; p = 0.51), reduced the carbon footprint of participants’ selections compared to control. The climate and climate + health swaps led to larger reductions in carbon footprint than the health swaps (ps < 0.001) but did not differ from one another (p = 0.43). Swaps’ effects on healthfulness and carbon footprint did not differ by age, health consciousness, or environmental consciousness (ps for interactions ≥0.10). The main limitations were that the sample differed somewhat from the US population and study results may not generalize to brick-and-mortar stores.
Author summary
Why was this study done?
- Modeling studies show that simple dietary substitutions (i.e., “swaps”) in which consumers replace one food with another similar food can improve the healthfulness of consumers’ diets and reduce the climate impact of food production.
- Retailers, including online grocery stores, have begun recommending health- or climate-focused food swaps, but little is known about how consumers respond to these swap recommendations in online grocery store settings.
What did the researchers do and find?
- This study used a randomized parallel trial design in a realistic online grocery store to evaluate the effects of recommending swaps to lower-climate-impact foods (“climate swaps”), more healthful foods (“health swaps”), or both (“climate + health swaps”).
- Participants received an average of 2.3–3.6 swap recommendations and accepted an average of 25%–27% of these swaps.
- The climate, health, and climate + health swaps all improved the healthfulness of participants’ grocery selections, but only the climate and climate + health swaps reduced the climate impact of selections.
What do these findings mean?
- Recommending swaps to more healthful or lower-climate-impact foods is a promising, scalable strategy for improving the healthfulness of online grocery purchases.
- Recommending swaps to lower-climate-impact foods may have the additional benefit of also reducing the climate impact of food production.
- The main limitations of the trial were that the sample differed somewhat from the US population overall and that study results may not generalize to brick-and-mortar stores.
Citation: Grummon AH, O’Sullivan K, Zeitlin AB, Taillie LS, Lee CJY, Walker AJ, et al. (2026) Effect of food swap recommendations on the healthfulness and carbon footprint of grocery purchases in a naturalistic online store: A randomized parallel trial. PLoS Med 23(8): e1004847. https://doi.org/10.1371/journal.pmed.1004847
Academic Editor: Jean Adams, University of Cambridge, UNITED KINGDOM OF GREAT BRITAIN AND NORTHERN IRELAND
Received: November 14, 2025; Accepted: July 30, 2026; Published: August 21, 2026
Copyright: © 2026 Grummon et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Individual-level survey and grocery selection data underlying this study are freely and publicly available at the following public repositories: https://doi.org/10.5281/zenodo.21420796 and https://github.com/rulm0000/swaps. Analytic code for this study is freely and publicly available at the following public repositories: https://doi.org/10.5281/zenodo.21420796 and https://github.com/rulm0000/swaps.
Funding: This research was supported by a grant from the National Institutes of Health, National Heart, Lung, and Blood Institute (K01 HL158608) to AHG (https://www.nhlbi.nih.gov/). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: I have read the journal’s policy and the authors of this manuscript have the following competing interests: LST is an Academic Editor on PLOS Medicine’s editorial board. The other authors have declared that no competing interests exist.
Introduction
Changing the foods we eat could improve health and mitigate climate change. Poor diet quality accounts for one in five deaths globally [1], and food production accounts for up to one-third of human-caused greenhouse gas emissions [2]. Experts agree that substantial dietary change is needed to reduce both diet-related disease and greenhouse gas emissions [3,4].
Many consumers report wanting to eat healthier, more environmentally sustainable diets [5–9]. However, consumers often struggle to realize these goals [10,11], in part because many find it hard to understand which foods promote health or have lower climate impacts (e.g., emit fewer greenhouse gases) [8,9]. Dietary guidelines, academic reports, and nongovernmental organizations recommend diets high in fruits, vegetables, and legumes and low in red and processed meats [3,12–17]. However, these broad recommendations may be challenging to implement given the complexity of food choice, including the need to consider multiple dimensions of food and compare many choices simultaneously [18–20].
One approach to addressing these challenges is to encourage consumers to make simple dietary substitutions, or “swaps,” in which they replace one food with another similar food [21]. Swaps could improve health and lower the climate impact of food production because foods that are otherwise similar (e.g., a beef burger and a turkey burger) can differ substantially in both their healthfulness and in the carbon emissions released when they are produced (i.e., their “carbon footprint”) [21,22]. Swaps may also be easier to implement than more extensive dietary changes like eliminating meat [23–25].
Technological advances have made it possible to recommend swaps to consumers in real time as they decide what foods to buy. For example, Kroger—the largest grocer in the US [26]—offers an online program called “OptUP” that recommends swaps from less healthful foods to similar, more healthful foods [27] and apps like “ecoSwitch” allow consumers to scan foods’ barcodes to receive swap recommendations to similar, lower-climate-impact foods [28,29]. Consumers are increasingly interested in receiving these recommendations [30].
Although simulation studies highlight the potential for swaps to improve the healthfulness or reduce the climate impact of food purchases [21,22,31–33], little is known about how consumers respond to swap recommendations. A small body of research indicates that recommending more healthful food swaps can improve the healthfulness of food purchases (e.g., reduce sodium purchases) [34–37], but studies have not tested whether recommending lower-climate-impact food swaps reduces the carbon footprint of purchases. Also unknown is whether recommending more healthful swaps influences carbon footprints and vice versa. Health- and climate-focused swaps might generate positive spillovers on the other dimension (i.e., a win-win), given the overlap between more healthful and lower-climate-impact dietary patterns [3,12,38], or negative spillovers, because some more healthful foods are relatively high-climate-impact (e.g., low-fat dairy products) and some lower-climate-impact foods are relatively unhealthful (e.g., meat mimics) [39]. Given these potential tradeoffs, swap recommendation systems may need to address both dimensions simultaneously, but the effects of doing so remain unknown.
To address these gaps, we tested whether recommending more healthful swaps, lower-climate-impact swaps, or both improves the healthfulness and reduces the carbon footprint of grocery purchases. We conducted a randomized trial in a naturalistic online grocery store. Food and beverage purchases in naturalistic stores have been shown to correlate with purchases from real stores [40,41].
Methods
Study design and participants
The study adopted a randomized parallel group design. In 2025, we recruited a convenience sample of adults through the survey research panel Prolific. The first participant was enrolled on May 7, 2025, and the last on June 4, 2025; follow-up was completed on June 25, 2025. Inclusion criteria were being 18 years of age or older, residing in the US, being able to complete a survey in English, and having internet access to complete 3 online study visits; exclusion criteria were being younger than 18, residing outside of the US, being unable to complete a survey in English, or not having internet access to complete 3 online survey visits. We used quota sampling to ensure that approximately 50% of participants were young adults ages 18–25 to enable moderation analyses by age group, given that age may influence intervention receptivity [42,43]. We followed CONSORT-2025 guidelines for reporting clinical trials [44] (S1 Checklist). Participants were not involved in the design, conduct, or reporting of the trial. The study was prospectively registered (ClinicalTrials.gov NCT#06648226). There were no deviations from the protocol.
Ethics.
Participants provided electronic written informed consent. The Stanford University Institutional Review Board approved the study (IRB#76925). There were no amendments to the IRB during the course of the trial.
Setting.
The trial took place in a naturalistic online grocery store developed for this study. The store displayed 5,550 products across 6 departments (S1 Table); to ensure the store closely mirrored the look and feel of a real online grocer, product images, names, prices, and nutritional information were displayed as they were shown on a large US grocery chain’s website in September 2023. The store’s functionalities also mirrored a real online store, including browsing, search, product pages, sorting and filtering, and checkout [40]. Online stores like this one are realistic, acceptable, and valid settings for evaluating nutrition interventions [40,45].
Carbon footprint of products.
We assessed the climate impact of all products in the store, operationalized as products’ carbon footprint (i.e., the greenhouse gas emissions associated with producing the products). We focused on carbon footprints because manufacturers and retailers including Unilever and Walmart have pledged to reduce carbon footprints, including from downstream consumption [46–50]. We obtained information on carbon footprints from GreenChoice, which calculated kg of carbon dioxide equivalents (CO2-eq) per kg for each item using a comprehensive review of farm-to-gate life cycle analyses [51].
Healthfulness of products.
We assessed product healthfulness using Ofcom Nutrient Profiling Model scores [52,53]. The Ofcom system was developed by the United Kingdom to regulate food marketing to children; scores are associated with risk of developing diet-related disease [54,55]. Products with more healthful Ofcom scores generally had lower carbon footprints, although the strength of this association varied across food groups (S1 Table).
Randomization and masking
Participants completed 3 online study visits. Participants were randomized at the end of the first study visit by Qualtrics to 1 of 4 trial arms in a 1:1:1:1 allocation ratio: climate swaps, health swaps, climate + health swaps, or control (Fig 1). Interventions are described below. Participants were not masked to assignments.
Procedures
Participants completed 3 online study visits, each approximately 1 week apart. At each visit, participants shopped in the naturalistic online store, then answered survey questions. In the first visit (baseline), participants provided electronic informed consent and shopped in the store with no intervention. Participants were randomized at the end of the first visit. In both the second and third visits, participants shopped in the store in their assigned arm; the third visit provided a second intervention exposure to enable us to test whether intervention effects changed with repeated exposure. At the start of each visit, participants completed a “commitment check” to encourage thoughtful responses [56].
When shopping in the store, participants were instructed to shop as they usually would for items in 6 food groups: beverages; boxed and frozen meals; milk and dairy products; meat and meat alternatives; soups; and sweets and snacks. These groups are top contributors to saturated fat, added sugar, sodium, and energy intake [57–62] or to the dietary carbon footprint in the US [21]. Participants were given a $50 budget based on national spending data [63] and 2022 sales data from the grocery chain on which the store was modeled. Participants could spend between 0.5 and 1.5 times this budget ($25 to $75) and otherwise shop freely. Participants were informed that 25 of them would be randomly chosen to receive the groceries selected in 1 of their visits as well as the remaining balance of their budget. This procedure incentivized participants to behave as they would in the real world, including selecting only items they actually wished to receive and considering whether they would rather have a particular item or retain the value of the item as cash. Participants were required to demonstrate understanding of this procedure before continuing. Because delivering groceries to a national sample of participants posed logistical challenges, we debriefed participants at the end of the study that those selected for delivery would instead receive an additional $50 incentive. Participants received $4.50, $6.00, and $8.00 USD for completing the first, second, and third visits, respectively.
Interventions.
In the climate swaps arm, the grocery store displayed color-coded climate grade labels (“A,” “B,” “C,” “D,” and “F”) next to all products, based on quintiles of carbon footprints in each food group (S1 Text). Grades were selected to match the predominant US grading system, which does not use the letter “E.” The labels mimicked Nutri-Score labels [64–66] (Fig 2A). When participants attempted to add a product with a “C,” “D,” or “F” grade, the store automatically offered them swaps to lower-climate-impact products (“A” or “B” grade) within the same product subcategory, category, or department as the originally selected product (see Fig 2B for an example swap and S2 Table for the hierarchy of subcategories, categories, and departments). S1 Text details how swaps were identified and presented; briefly, the store identified all possible swaps (i.e., products with an “A” or “B” grade in the same subcategory, category, or department as the originally selected product) and presented an initial set of up to 8 products, beginning with those in the same subcategory as the originally selected product. Participants could load additional swaps, in batches of up to 8, until all possible swaps were shown. Climate swap recommendations were based solely on carbon footprint.
Note: Figure shows climate labels and an example climate swap. The health labels had the same design as the climate labels but used the word “Nutrition” instead of “Climate.” Panel B shows mock products to preserve copyright; participants saw real, branded products in the trial.
In the health swaps arm, the online grocery store displayed health grade labels based on Ofcom scores [52,53], using the same label design as above. We applied “A” or “B” grades to products above the United Kingdom’s regulatory cutoffs for marketing products to children [53] and “C,” “D,” and “F” grades to products below these cutoffs (S1 Text). When participants attempted to add a product with a “C,” “D,” or “F” grade, the store automatically offered swaps to similar but more healthful products (with an “A” or “B” grade) that were in the same product subcategory, category, or department as the originally selected product (S1 Text). Health swap recommendations were based solely on healthfulness.
In the climate + health swaps arm, the store displayed both the climate and health labels side-by-side next to all products. When participants attempted to add a product with a “C,” “D,” or “F” grade on either dimension, the store offered swaps that improved at least 1 dimension (carbon footprint or healthfulness) and were at least as good or better on the other, while never recommending products with a “C,” “D,” or “F” grade on either dimension (S1 Text).
Outcomes
The 2 prespecified primary endpoints were the healthfulness and carbon footprint of participants’ selections. Healthfulness was defined as the average Ofcom score of selected products, converted to a 1–100 scale (higher scores indicate more healthful selections) [53]. Carbon footprint was defined as the mean greenhouse gas emissions intensity (i.e., kg of carbon dioxide equivalents [CO2-eq] per kg) of selected products. The study evaluated several secondary endpoints, including both selection endpoints (related to participants’ grocery selections) and psychological endpoints (assessed via survey). Secondary selection endpoints were energy (i.e., calorie) and nutrient (i.e., sugar, sodium, saturated fat, fiber, and protein) density of selections and total spending. Secondary psychological outcomes were selected based on health behavior theory and previous research [67–73]. These outcomes were thinking about health, climate impact, and taste while shopping (also called “cognitive elaboration”); feeling negative (i.e., guilt, shame, worry) and positive (i.e., inspiration, pride, reassurance) emotions while shopping; and injunctive and descriptive norms (i.e., beliefs about what people should do and what they typically do, respectively) about buying healthy foods and foods with low climate impact. As an exploratory endpoint, the survey also assessed acceptability of the labels and swaps [74].
Finally, the survey assessed participants’ characteristics, including demographics, health consciousness [75,76], and environmental consciousness [77], and their reactions to the store (S3 Table shows all measures). All outcomes were pre-registered prior to data collection and are presented in the current manuscript. Harms were assessed by participant report.
Statistical analysis
We pre-registered the analysis plan (ClinicalTrials.gov NCT#06648226; S1 Protocol); deviations to this plan are noted. Analyses were intent-to-treat and used a two-tailed alpha = 0.05. We conducted analyses in Stata MP version 19.
First, we summarized the number of swaps offered and accepted. Second, we used linear mixed models to estimate the effects of the 3 swaps interventions on the 2 primary outcomes, pooling across the first and second exposure to maximize power to detect effects (S1 Text). Mixed effects models handled missing data using maximum likelihood. To estimate within-person change from baseline, we regressed the outcome on indicator variables for each trial arm (i.e., climate swaps, health swaps, and climate + health swaps, treating the control group as the reference category; this adjusts for baseline values of the outcome), study visit (treating the baseline visit as the reference category), the interaction between trial arm and visit, and a random intercept to account for repeated measures within participants. We did not include any covariates given the randomized design. We used the models to estimate changes from baseline to the intervention period (weighted average of visits 2 and 3) relative to control and to test differences between the interventions. We estimated corrected p-values using the Bonferroni-Holm method [78], considering 6 tests for the primary outcomes (each intervention arm versus control for the 2 primary outcomes) and 3 tests for the secondary outcomes (each intervention arm versus control); these corrections were not pre-registered. This change did not affect the statistical significance of the results overall, except for two tests examining thinking about health. We converted effect sizes to standardized effects (Cohen’s d) using the Campbell Collaboration’s Effect Size Calculator [79], interpreting effects of d = .20 as small, .50 as medium, and .80 as large [80].
Third, we examined intervention effects on the primary outcomes within food groups (e.g., separately for beverages, meals, etc.); these analyses were exploratory and not pre-registered. Fourth, we examined whether the effects of the swaps interventions on the 2 primary outcomes were moderated by age (young adult [18–25 years] versus older adult [≥26 years]), health consciousness (treated continuously), and environmental consciousness (treated continuously) by adding indicator variables for the moderator and interactions between the moderator, trial arm, and study visit to the main model. We used separate models for each moderator and tested the joint significance of the three-way interaction terms. We chose these moderators because they may influence receptivity to the interventions [42,43,77,81]. Fifth, we examined whether intervention effects on the 2 primary outcomes changed from the first to the second exposure by examining the contrast between effects at the first versus second exposure. Sixth, we examined effects on secondary outcomes using the same mixed effects approach. Seventh, we described intervention acceptability (e.g., proportion who liked the swaps).
We estimated a priori sample size needs to detect an effect of each intervention versus control of d = .15 or larger, a small effect [80] that is conservative based on prior studies [34,82,83]. A priori power calculations assumed a two-tailed alpha = 0.05 and within-person correlation = .40, and 90% power. Under these assumptions, a sample of 904 (226 per group) would yield 90% power to detect an effect of this size or larger. To allow for missing data, we aimed to recruit 1,200 participants (300 per group). We ultimately included 1,201 participants in analyses; assuming a more conservative alpha of 0.008 (0.05 ÷ 6 tests; more conservative than the Bonferroni-Holm correction we applied post hoc), this achieved sample size yielded 87% power to detect effects of d = .15 or larger.
Artificial intelligence tools and technologies
We used ChatGPT 4o-mini to create the mock product images shown in Fig 2B. We show mock products in the figure to preserve copyright; participants saw real, branded products in the trial. No other content in the manuscript or supporting files was generated by artificial intelligence tools or technologies.
Results
Sample characteristics
A total of 1,201 participants were randomized and included in analyses (Fig 1). No adverse events were reported. Participants’ mean [SD] age was 31.5 [12.2] years and 53% (638/1,201) identified as women (Table 1). About two-thirds (67%; 810/1,201) reported annual household income <$100,000. Compared with the US population, participants were more likely to be young adults (consistent with recruitment goals), identify as Black or African American, have a college or graduate degree, and have income <$100,000, and were less likely to identify as Latino or Hispanic or multiracial (S4 Table).
Acceptability of the online store
Participants found the online store realistic and acceptable (S1 Fig). For example, 94% (1,133/1,201) reported they would pick similar products in the real world as they selected in the online store.
Swaps offered and accepted
Participants selected a mean [SD] of 11.3 items [4.0]. Pooling across intervention visits, participants in the climate swaps arm were offered a mean [SD] of 2.4 [2.8] swaps, those in the health swaps arm 2.3 [2.7] swaps, and those in the climate + health swaps arm 3.6 [3.8] swaps. On average, participants in the climate (25%), health (27%), and climate + health (27%) swaps arms accepted a similar proportion of the swaps offered.
Healthfulness of grocery selections
At baseline (no intervention), participants selected groceries with a mean [SD] healthfulness of 57.7 [8.2] on the 1–100 scale. Compared with control, the climate (B = 2.4, 95% CI [1.1, 3.7]), health (B = 4.7, 95% CI [3.3, 6.0]), and climate + health (B = 4.0, 95% CI [2.7, 5.4]) swaps all improved the healthfulness (all corrected ps < 0.001, Table 2, Fig 3). The health (difference versus climate swaps = 2.3, 95% CI [0.9, 3.6]; p < 0.001) and climate + health (difference versus climate swaps = 1.6, 95% CI [0.3, 3.0]; p = 0.02) swaps both increased the overall healthfulness of selections more than climate swaps alone but did not differ from one another (p = 0.34). The benefits of the health and climate + health swaps on healthfulness were observed for most food groups, whereas the benefits of the climate swaps on healthfulness were observed for only two food groups, meat and meat alternatives and sweets and snacks (S5 Table). The effects of the swaps on healthfulness did not differ by age group (p for joint significance of interaction terms = 0.17), health consciousness (p = 0.97), or environmental consciousness (p = 0.41).
Note: Figure shows raw means and standard errors (SEs).
The effects of the health and climate + health swaps on the healthfulness of food selections increased somewhat after a second exposure (visit 3 versus visit 2, differences = 1.3 and 1.2, respectively), but these differences were not statistically significant (ps = 0.11 and 0.13, respectively) (Fig 3). The effects of the climate swaps on healthfulness did not change across exposures (p = 0.94).
Carbon footprint of grocery selections
At baseline, participants selected groceries with mean [SD] carbon footprint of 8.5 kg [7.1] CO2-eq per kg. Compared with control, the climate (B = −3.3, 95% CI [−4.6, −2.1]; corrected p < 0.001; approximate 39% reduction) and climate + health swaps (B = −2.8, 95% CI [−4.1, −1.6]; corrected p < 0.001; approximate 33% reduction), but not the health swaps (B = −0.4, 95% CI [−1.7, 0.8]; corrected p = 0.51), reduced the carbon footprint of participants’ selections. Both the climate (difference versus health swaps = −2.9, 95% CI [−4.2, −1.7]; p < 0.001) and climate + health (difference versus health swaps = −2.4, 95% CI [−3.7, −1.2]; p < 0.001) swaps produced larger reductions in carbon footprint overall than the health swaps but did not differ from one another (p = 0.43). The benefits of the climate swaps on carbon footprints were observed across all food groups except beverages, whereas the benefits of the climate + health swaps were observed only for meat and meat alternatives, soups, and sweets and snacks (S5 Table). Although the health swaps did not reduce carbon footprint overall, they reduced the carbon footprint of milk and dairy selections and sweets and snacks selections (S5 Table). The effects of the swaps on carbon footprint did not differ by age group (p for joint significance of interactions = 0.34), health consciousness (p = 0.10), or environmental consciousness (p = 0.35). Additionally, the effects of the swaps on the carbon footprint of participants’ grocery selections did not change across exposures (ps > .39, Fig 3).
Energy and nutrient densities and total spending
All three swap interventions (climate, health, and climate + health) reduced the energy and saturated fat density of selections compared to control (corrected ps < 0.05; Table 2). The health and climate + health swaps also reduced the sugar density of selections. Only the health swaps increased the fiber density of selections, and none of the interventions affected sodium or protein density. All three interventions reduced total spending (corrected ps < 0.01).
Psychological responses
The health swaps increased how much participants reported thinking about health while shopping (corrected ps < 0.05), but the climate and climate + health swaps did not (both corrected ps = 0.07) (Table 2). Both the climate and climate + health swaps (both corrected ps < 0.001), but not the health swaps, increased thinking about climate impact. None of the interventions affected thinking about taste. All three interventions increased negative emotions while shopping (all corrected ps ≤ 0.01), and none affected positive emotions, injunctive norms, or descriptive norms.
Acceptability of the interventions
Participants perceived the interventions as acceptable. Across trial arms, a large majority of participants somewhat or strongly approved of grocery stores implementing the climate (77%, 791/1,030) and health (85%, 873/1,030) swaps and the accompanying climate (78%, 805/1,030) and health (87%, 891/1,030) labels. Most also indicated that the swaps and labels would help them quite a bit or a lot in choosing lower-climate-impact or healthier foods and that they liked the swaps and labels quite a bit or a lot (Fig 4).
Discussion
In this randomized trial, climate and health swap recommendations, alone and in combination with one another, improved the healthfulness of grocery selections. Although improvements appear modest (approximately 2–5 points on the Ofcom scale), a prospective cohort study found that each 2-point increase in diet quality as assessed by Ofcom scores was associated with 14% lower risk of developing cardiovascular disease [55]. The climate swaps, alone or in combination with health swaps, also reduced the average carbon footprint of grocery selections by approximately 33%–39%, whereas the health swaps alone did not. Several major retailers already recommend swaps to their customers [27–29]; these findings suggest that such recommendations—whether focused on climate, health, or both—are a promising strategy to improve the healthfulness of grocery purchases and that climate-focused swaps may additionally reduce food-related greenhouse gas emissions.
The climate swaps improved the healthfulness of grocery selections even when offered without health swaps. By contrast, the health swaps alone did not reduce the carbon footprint of selections. The climate swaps also led participants to think more about health while shopping, while the health swaps did not increase thinking about climate impact. These results suggest asymmetric halo effects in which consumers generalize to perceive lower-climate-impact products to be more healthful, but do not generalize the other direction to perceive more healthful products as having lower climate impact. Although halo effects can sometimes undermine healthy or sustainable decision-making [84], the lower-climate-impact products offered in our store were, on average, more healthful, suggesting that generalizing was not unreasonable. Climate-focused interventions may therefore serve as a “stealth” strategy for encouraging more healthful purchases [85], particularly among consumers less motivated by health-focused approaches.
The health swaps improved the healthfulness of grocery selections more than the climate swaps, whereas the climate swaps reduced the carbon footprint of selections more than the health swaps. Combining the health and climate swaps did not dilute the effects of either alone: the climate + health swaps improved healthfulness to a similar extent as the health swaps alone and reduced carbon footprints to a similar extent as the climate swaps alone. These findings suggest that combining climate and health swaps does not confuse or overwhelm consumers or cause them to ignore both swaps or prioritize only one dimension (climate or health). Prior studies of combined climate and health labels have reported similar results [86,87]. Together, this evidence indicates that combining climate and health swaps may be a particularly promising strategy for addressing both diet-related disease and dietary carbon emissions. Moreover, both types of swaps were highly acceptable to participants.
The effects of the swaps did not vary by age group, health consciousness, or environmental consciousness. These results are consistent with studies of climate labels on restaurant menus [88,89] and taxes and warnings on red meat [73], which likewise found no differences in intervention effects by these characteristics. Together, this evidence suggests that swap recommendations may work equally well across a range of population groups.
Participants were exposed to the swaps twice, approximately 1 week apart. For participants in the health or climate + health swaps arms, we observed a modest but non-significant increase in healthfulness of selections after the second exposure compared to the first. This pattern is consistent with observational studies finding that the impact of graphic cigarette warnings may grow over time [90,91], though other research reports that warning effects attenuate over time [92]. By contrast, we observed no additional reductions in carbon footprint after the second exposure to the climate or climate + health swaps. Although longer studies will be needed, these results suggest that the benefits of health and climate swaps do not diminish rapidly and that the benefits of health swaps may increase modestly with repeated exposures.
Few studies have examined how swap recommendations exert their effects on behavior. In this study, the health swaps led consumers to think more about health while they were shopping (a process sometimes referred to as “cognitive elaboration”), and the climate swaps (alone or in combination with the health swaps) led consumers to think more about climate impact. All three interventions also increased negative emotions such as guilt and worry. Some scholars describe these negative emotional reactions as an “emotional tax” that should be weighed alongside intervention benefits [93,94], whereas others suggest that negative emotions can be an appropriate and meaningful response to certain interventions. Indeed, information that fails to elicit emotion often lacks meaning and is weighed less heavily in decision-making, potentially resulting in worse decisions [95–97]. Either way, our results suggest that cognitive elaboration and negative emotions may help explain how swap recommendations influence grocery purchases.
Participants in the intervention arms were offered approximately 2–4 swaps, indicating that only approximately 20%–30% of their original selections received “C,” “D,” or “F” grades. These results suggest that health and climate labels alone may have encouraged participants to select healthful or low-climate-impact items even before they were offered swaps. However, because labels and swaps were always delivered together in our study, we cannot disentangle their independent effects.
Strengths of this study included the use of a randomized controlled design, naturalistic online grocery store, and repeated exposures to the interventions. Limitations include using a convenience sample that differed somewhat from the overall US population, though a large body of evidence finds that trials in convenience samples typically produce similar results to representative samples [98–102]. Second, participants shopped in a simulated online grocery store (not an actual store) and did not pay for or receive their selections. However, they were incentivized to choose items they actually wanted, and prior research shows good correspondence between selections in simulated online grocery stores and real-world purchases [40]. Third, although the online store included more than 5,000 products, some food groups were not represented, so the effects on selections in those groups remain unknown. Fourth, we addressed price by having prices be the same across arms and designing the study to incentivize participants to consider cost as they would in the real world. However, we did not design swaps so that recommended products were similar in price to originally selected products. Swap acceptance may be higher when swaps are similarly priced to originally selected products; future studies could evaluate such price-neutral swaps. Fifth, we tested swaps in the context of an online grocery store, so effects in brick-and-mortar stores remain unknown. Sixth, we tested a combined climate + health intervention that presented both labels side-by-side and recommended swaps based on both labels. However, some retailers integrate climate and health information into a single score and display labels and swaps based on this integrated score; our results may not generalize to these systems. Seventh, data limitations precluded us from examining why participants accepted certain swaps and rejected others; future studies could explore what makes swaps acceptable to consumers, for example using qualitative approaches.
In summary, this randomized trial found that recommending swaps to more healthful or lower-climate-impact foods increases the healthfulness of grocery selections and recommending swaps to lower-climate-impact foods also reduces the carbon footprint of grocery selections. Combining both types of swaps improved both outcomes without diluting the impact of either intervention alone. Offering combined health- and climate-focused swaps in online grocery stores represents a promising, scalable strategy to reduce diet-related disease and the environmental impacts of food consumption.
Supporting information
S1 Table. Products shown in the naturalistic online store used in the trial.
https://doi.org/10.1371/journal.pmed.1004847.s001
(DOCX)
S2 Table. Departments, categories, and subcategories used in the naturalistic online store used in the trial.
https://doi.org/10.1371/journal.pmed.1004847.s002
(DOCX)
S4 Table. Comparison of characteristics of the study sample (n = 1,201 US adults) to national estimates.
https://doi.org/10.1371/journal.pmed.1004847.s004
(DOCX)
S5 Table. Effects of the climate, health, and climate + health swaps on the healthfulness and carbon footprint of food and beverage purchases, by food group, n = 1,201 US adults.
https://doi.org/10.1371/journal.pmed.1004847.s005
(DOCX)
S1 Fig. Acceptability of the naturalistic online store used in the trial.
https://doi.org/10.1371/journal.pmed.1004847.s007
(DOCX)
S1 Protocol. Approved IRB Protocol and Statistical Analysis Plan.
https://doi.org/10.1371/journal.pmed.1004847.s008
(PDF)
S1 Checklist. CONSORT-2025 Checklist.
Hopewell S, Chan AW, Collins GS, Hróbjartsson A, Moher D, Schulz KF, et al. CONSORT 2025 Statement: updated guideline for reporting randomised trials. BMJ. 2025; 388:e081123. https://dx.doi.org/10.1136/bmj-2024-081123. 2025 Hopewell and colleagues. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
https://doi.org/10.1371/journal.pmed.1004847.s009
(DOCX)
Acknowledgments
We thank Sara Cathey for work developing the product labels and the logo for the online grocery store.
References
- 1. Afshin A, Sur PJ, Fay KA, Cornaby L, Ferrara G, Salama JS. Health effects of dietary risks in 195 countries, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. The Lancet. 2019;393:1958–72.
- 2. Crippa M, Solazzo E, Guizzardi D, Monforti-Ferrario F, Tubiello FN, Leip A. Food systems are responsible for a third of global anthropogenic GHG emissions. Nat Food. 2021;2(3):198–209. pmid:37117443
- 3. Willett W, Rockström J, Loken B, Springmann M, Lang T, Vermeulen S, et al. Food in the Anthropocene: the EAT-Lancet Commission on healthy diets from sustainable food systems. Lancet. 2019;393(10170):447–92. pmid:30660336
- 4. Springmann M, Clark M, Mason-D’Croz D, Wiebe K, Bodirsky BL, Lassaletta L, et al. Options for keeping the food system within environmental limits. Nature. 2018;562(7728):519–25. pmid:30305731
- 5.
Leiserowitz A, Ballew M, Rosenthal S, Semaan J. Climate change and the American diet. New Haven, CT: Yale University and EarthDay Network; 2020.
- 6. Buchholz K. Infographic: are Americans trying to eat healthy? In: Statista Daily Data [Internet]. 24 Feb 2023 [cited 26 Feb 2024]. Available from: https://www.statista.com/chart/16796/us-interest-in-healthy-food
- 7. Sodexo. Sodexo and Harris interactive introduce action-oriented international sustainable food barometer. In: Sodexo USA [Internet]. 16 Jan 2024 [cited 26 Feb 2024]. Available from: https://us.sodexo.com/home/media/news-releases/newsList-area/press-releases/sustainable-food-barometer.html
- 8. McKinsey & Co. Good eating. In: McKinsey & Company [Internet]. 1 May 2023 [cited 10 June 2024]. Available from: https://www.mckinsey.com/featured-insights/sustainable-inclusive-growth/chart-of-the-day/good-eating
- 9.
Grimmelt A, Moulton J, Pandya C, Snezhkova N. Hungry and confused: the winding road to conscious eating. McKinsey & Company. 2022.
- 10. Hill JO, Wyatt HR, Reed GW, Peters JC. Obesity and the environment: where do we go from here?. Science. 2003;299(5608):853–5. pmid:12574618
- 11. Roberto CA, Swinburn B, Hawkes C, Huang TT-K, Costa SA, Ashe M, et al. Patchy progress on obesity prevention: emerging examples, entrenched barriers, and new thinking. Lancet. 2015;385(9985):2400–9. pmid:25703111
- 12. Clark MA, Springmann M, Hill J, Tilman D. Multiple health and environmental impacts of foods. Proc Natl Acad Sci U S A. 2019;116(46):23357–62. pmid:31659030
- 13. Poore J, Nemecek T. Reducing food’s environmental impacts through producers and consumers. Science. 2018;360(6392):987–92. pmid:29853680
- 14. Nelson ME, Hamm MW, Hu FB, Abrams SA, Griffin TS. Alignment of healthy dietary patterns and environmental sustainability: a systematic review. Adv Nutr. 2016;7(6):1005–25. pmid:28140320
- 15. Van Horn L, Carson JAS, Appel LJ, Burke LE, Economos C, Karmally W, et al. Recommended dietary pattern to achieve adherence to the American Heart Association/American College of Cardiology (AHA/ACC) guidelines: a scientific statement from the American Heart Association. Circulation. 2016;134(22):e505–29. pmid:27789558
- 16. Willits-Smith A, Aranda R, Heller MC, Rose D. Addressing the carbon footprint, healthfulness, and costs of self-selected diets in the USA: a population-based cross-sectional study. Lancet Planet Health. 2020;4(3):e98–106. pmid:32220679
- 17. Gonzalez Fischer C, Garnett T. Plates, pyramids, planet. Developments in national healthy and sustainable dietary guidelines: a state of play assessment. In: Food and Agriculture Organization of the United Nations and Food Climate Research Network [Internet]. Food and Agriculture Organization of the United Nations and Food Climate Research Network; 2016 [cited 11 Feb 2026]. Available from: https://www.fao.org/3/a-i5640e.pdf
- 18. Leng G, Adan RAH, Belot M, Brunstrom JM, de Graaf K, Dickson SL, et al. The determinants of food choice. Proc Nutr Soc. 2017;76(3):316–27. pmid:27903310
- 19. Sobal J, Bisogni CA, Devine CM, Jastran M. A conceptual model of the food choice process over the life course. The psychology of food choice. CABI. 2006. p. 1–18.
- 20. Köster EP. Diversity in the determinants of food choice: a psychological perspective. Food Qual Prefer. 2009;20(2):70–82.
- 21. Grummon AH, Lee CJY, Robinson TN, Rimm EB, Rose D. Simple dietary substitutions can reduce carbon footprints and improve dietary quality across diverse segments of the US population. Nat Food. 2023;4(11):966–77. pmid:37884673
- 22. Rose D, Willits-Smith AM, Heller MC. Single-item substitutions can substantially reduce the carbon and water scarcity footprints of US diets. Am J Clin Nutr. 2022;115(2):378–87. pmid:35024805
- 23. Bandura A. Health promotion from the perspective of social cognitive theory. Psychol Health. 1998;13(4):623–49.
- 24. Hills AP, Byrne NM, Lindstrom R, Hill JO. “Small changes” to diet and physical activity behaviors for weight management. Obes Facts. 2013;6(3):228–38. pmid:23711772
- 25. Hill JO. Can a small-changes approach help address the obesity epidemic? A report of the Joint Task Force of the American Society for Nutrition, Institute of Food Technologists, and International Food Information Council. Am J Clin Nutr. 2009;89(2):477–84. pmid:19088151
- 26. Kantar and the National Retail Foundation. Leading supermarkets in the United States in 2022, based on retail sales. In: Statista [Internet]. July 2023 [cited 7 Oct 2024]. Available from: https://www.statista.com/statistics/197899/2010-sales-of-supermarket-chains-in-the-us/
- 27. Kroger. OptUP FAQs. In: Kroger [Internet]. 2024 [cited 23 Mar 2024]. Available from: https://www.kroger.com/hc/help/faqs/health-and-wellness/optup
- 28. Wells J. Personalized health programs are coming to grocery stores. In: Grocery Dive [Internet]. 25 Apr 2022 [cited 22 Mar 2024]. Available from: https://www.grocerydive.com/news/heinens-grocery-personalized-health-programs/622442/
- 29. The George Institute. Power in your hands: FoodSwitch and ecoSwitch. In: The George Institute for Global Health [Internet]. 2024 [cited 22 Mar 2024]. Available from: https://www.georgeinstitute.org/projects/foodswitch
- 30. Sweeney J. Shoppers are looking for “healthy swaps.” In: Grocery Dive [Internet]. 28 Feb 2019 [cited 22 Mar 2024]. Available from: https://www.grocerydive.com/news/shoppers-are-looking-for-healthy-swaps/549395/
- 31. Auclair O, Eustachio Colombo P, Milner J, Burgos SA. Partial substitutions of animal with plant protein foods in Canadian diets have synergies and trade-offs among nutrition, health and climate outcomes. Nat Food. 2024;5(2):148–57. pmid:38365909
- 32. O’Sullivan K, Tiefenbeck V, Stucki M, Bradford S, Fuchs KL, Wanner S. A single specific and actionable swap recommendation can lead to substantial reductions in households’ dietary carbon footprints. J Clean Prod. 2025;526:146577.
- 33. Gaines A, Shahid M, Coyle D, Barrett E, Hadjikakou M, Wu JHY, et al. Switches in food and beverage product purchases can reduce greenhouse gas emissions in Australia. Nat Food. 2024;5(6):524–32. pmid:38806695
- 34. Koutoukidis DA, Jebb SA, Ordóñez-Mena JM, Noreik M, Tsiountsioura M, Kennedy S, et al. Prominent positioning and food swaps are effective interventions to reduce the saturated fat content of the shopping basket in an experimental online supermarket: a randomized controlled trial. Int J Behav Nutr Phys Act. 2019;16(1):50. pmid:31174547
- 35. Forwood SE, Ahern AL, Marteau TM, Jebb SA. Offering within-category food swaps to reduce energy density of food purchases: a study using an experimental online supermarket. Int J Behav Nutr Phys Act. 2015;12:85. pmid:26109390
- 36. Piernas C, Lee C, Hobson A, Harmer G, Payne Riches S, Noreik M, et al. A behaviorally informed mobile app to improve the nutritional quality of grocery shopping (SwapSHOP): feasibility randomized controlled trial. JMIR Mhealth Uhealth. 2024;12:e45854. pmid:38206671
- 37. Hartmann-Boyce J, Bianchi F, Piernas C, Payne Riches S, Frie K, Nourse R, et al. Grocery store interventions to change food purchasing behaviors: a systematic review of randomized controlled trials. Am J Clin Nutr. 2018;107(6):1004–16. pmid:29868912
- 38. Tilman D, Clark M. Global diets link environmental sustainability and human health. Nature. 2014;515(7528):518–22. pmid:25383533
- 39. Potter C, Bastounis A, Hartmann-Boyce J, Stewart C, Frie K, Tudor K, et al. The effects of environmental sustainability labels on selection, purchase, and consumption of food and drink products: a systematic review. Environ Behav. 2021;53(8):891–925. pmid:34456340
- 40. Grummon AH, Tucker AC, Noe V, Rummo PE, Prestemon CE, Hall MG, et al. Consumer behaviour and experiences in a naturalistic online grocery store: implications for nutrition research. J Nutr Sci. 2023;12:e36. pmid:37008411
- 41. Hall MG, Higgins ICA, Grummon AH, Lazard AJ, Prestemon CE, Sheldon JM, et al. Using a naturalistic store laboratory for clinical trials of point-of-sale nutrition policies and interventions: a feasibility and validation study. Int J Environ Res Public Health. 2021;18(16):8764. pmid:34444509
- 42. Bunten A, Porter L, Sanders JG, Sallis A, Payne Riches S, Van Schaik P, et al. A randomised experiment of health, cost and social norm message frames to encourage acceptance of swaps in a simulation online supermarket. PLoS One. 2021;16(2):e0246455. pmid:33596251
- 43. LaRose JG, Leahey TM, Hill JO, Wing RR. Differences in motivations and weight loss behaviors in young adults and older adults in the National Weight Control Registry. Obesity (Silver Spring). 2013;21(3):449–53. pmid:23404944
- 44. Hopewell S, Chan A-W, Collins GS, Hróbjartsson A, Moher D, Schulz KF. CONSORT 2025 statement: updated guideline for reporting randomized trials. JAMA. 2025.
- 45. Rummo PE, Higgins I, Chauvenet C, Vesely A, Jaacks LM, Taillie L. A standardized guide to developing an online grocery store for testing nutrition-related policies and interventions in an online setting. Int J Environ Res Public Health. 2021;18(9):4527. pmid:33923246
- 46. Casey C. Food companies tackling Scope 3 emissions despite weak SEC rule: expert. In: Food Dive [Internet]. 18 Mar 2024 [cited 6 Apr 2024]. Available from: https://www.fooddive.com/news/sec-climate-rule-scope-3-emissions-carbon-footprint-pepsico-howgood-tracking/708947/
- 47. Jesse B, Perotti A, Roos D. Decarbonizing grocery | McKinsey. McKinsey and Company; 22 July 2022 [cited 11 Feb 2026]. Available from: https://www.mckinsey.com/industries/retail/our-insights/decarbonizing-grocery
- 48. Coyne A. The road to net zero – Big Food’s emissions pledges. In: Just Food [Internet]. 14 Mar 2024 [cited 6 Apr 2024]. Available from: https://www.just-food.com/features/the-road-to-net-zero-big-foods-emission-pledges/
- 49. Mirza Z. Unilever updates climate action plan, targets scope 3 emissions reduction. In: ESG Dive [Internet]. Mar 2024 [cited 6 Apr 2024]. Available from: https://www.esgdive.com/news/unilever-updates-climate-action-plan-targets-scope-3/710053/
- 50. Walmart Policies and Guidelines. In: Walmart Policies and Guidelines [Internet]. [cited 6 Apr 2024]. Available from: https://corporate.walmart.com/policies
- 51. GreenChoice. Food climate footprint score methodology. 2025 [cited 20 July 2025]. Available from: https://about.greenchoicenow.com/resources/food-climate-footprint-score-methodology
- 52. United Kingdom Department of Health. Nutrient Profiling Technical Guidance. United Kingdom Department of Health; 2011 [cited 12 June 2022]. Available from: https://www.gov.uk/government/publications/nutrient-profiling-model-2018/nutrient-profiling-model-2018-technical-guidance
- 53. Rayner M, Scarborough P, Lobstein T. The UK Ofcom Nutrient Profiling Model: defining “healthy” and “unhealthy” foods and drinks for TV advertising to children. 2009 [cited 12 June 2022]. Available from: https://docs.google.com/viewer?url= https%3A%2F%2Fwww.ndph.ox.ac.uk%2Fcpnp%2Ffiles%2Fabout%2Fuk-ofcom-nutrient-profile-model.pdf
- 54. Egnell M, Seconda L, Neal B, Mhurchu CN, Rayner M, Jones A, et al. Prospective associations of the original Food Standards Agency nutrient profiling system and three variants with weight gain, overweight and obesity risk: results from the French NutriNet-Santé cohort. Br J Nutr. 2021;125(8):902–14. pmid:32878658
- 55. Adriouch S, Julia C, Kesse-Guyot E, Méjean C, Ducrot P, Péneau S, et al. Prospective association between a dietary quality index based on a nutrient profiling system and cardiovascular disease risk. Eur J Prev Cardiol. 2016;23(15):1669–76. pmid:27000099
- 56. Geisen E. Improve data quality by using a commitment request instead of attention checks. In: Qualtrics [Internet]. 4 Aug 2022 [cited 11 Feb 2026]. Available from: https://www.qualtrics.com/blog/attention-checks-and-data-quality/
- 57. CDC. Sodium and Health. In: Centers for Disease Control and Prevention [Internet]. 6 June 2023 [cited 6 Apr 2024]. Available from: https://www.cdc.gov/salt/index.htm
- 58. Ahmed M, Ng AP, Christoforou A, Mulligan C, L’Abbé MR. Top sodium food sources in the American diet – using National Health and Nutrition Examination Survey. Nutrients. 2023;15(4):831. pmid:36839189
- 59. Wambogo EA, Ansai N, Terry A, Fryar C, Ogden C. Dairy, meat, seafood, and plant sources of saturated fat: United States, ages two years and over, 2017–2020. J Nutr. 2023;153:2689–98.
- 60.
United States Department of Agriculture and Department of Health and Human Services. Dietary guidelines for Americans, 2020-2025. U.S. Government Printing Office; 2020. Available from: http://www.dietaryguidelines.gov
- 61. Bailey RL, Fulgoni VL, Cowan AE, Gaine PC. Sources of added sugars in young children, adolescents, and adults with low and high intakes of added sugars. Nutrients. 2018;10(1):102. pmid:29342109
- 62. Huth PJ, Fulgoni VL, Keast DR, Park K, Auestad N. Major food sources of calories, added sugars, and saturated fat and their contribution to essential nutrient intakes in the US diet: Data from the National Health and Nutrition Examination Survey (2003–2006). Nutr J. 2013;12:1–10.
- 63. Bureau of Labor Statistics. Table R-1. All consumer units: Annual detailed expenditure means, standard errors, relative standard errors, and weekly (D) or quarterly (I) percents reporting, Consumer Expenditure Surveys, 2022. Bureau of Labor Statistics; 2022 [cited 11 Feb 2026]. Available from: https://www.bls.gov/cex/tables/calendar-year/mean/cu-all-detail-2022.pdf
- 64. Egnell M, Galan P, Fialon M, Touvier M, Péneau S, Kesse-Guyot E, et al. The impact of the Nutri-Score front-of-pack nutrition label on purchasing intentions of unprocessed and processed foods: post-hoc analyses from three randomized controlled trials. Int J Behav Nutr Phys Act. 2021;18(1):38. pmid:33731145
- 65. Egnell M, Boutron I, Péneau S, Ducrot P, Touvier M, Galan P, et al. Front-of-pack labeling and the nutritional quality of students’ food purchases: a 3-arm randomized controlled trial. Am J Public Health. 2019;109(8):1122–9. pmid:31219721
- 66. Talati Z, Egnell M, Hercberg S, Julia C, Pettigrew S. Food choice under five front-of-package nutrition label conditions: an experimental study across 12 countries. Am J Public Health. 2019;109(12):1770–5. pmid:31622139
- 67. Brewer NT, Parada H, Hall MG, Boynton MH, Noar SM, Ribisl KM. Understanding why pictorial cigarette pack warnings increase quit attempts. Ann Behav Med. 2019;53(3):232–43. pmid:29850764
- 68. Grummon AH, Brewer NT. Health warnings and beverage purchase behavior: mediators of impact. Ann Behav Med. 2020;54(9):691–702. pmid:32182336
- 69. Hall MG, Grummon AH, Queen T, Lazard AJ, Higgins ICA, Richter APC, et al. How pictorial warnings change parents’ purchases of sugar-sweetened beverage for their children: mechanisms of impact. Int J Behav Nutr Phys Act. 2023;20(1):76. pmid:37353823
- 70. Cacioppo JT, Petty RE. The elaboration likelihood model of persuasion. ACR North Am Adv. 1984.
- 71. Ajzen I. The theory of planned behavior. Organ Behav Hum Decis Process. 1991;50:179–211.
- 72. Cialdini RB, Kallgren CA, Reno RR. A focus theory of normative conduct: a theoretical refinement and reevaluation of the role of norms in human behavior. Advances in Experimental Social Psychology. Elsevier. 1991. p. 201–34.
- 73. Taillie LS, Bercholz M, Prestemon CE, Higgins ICA, Grummon AH, Hall MG, et al. Impact of taxes and warning labels on red meat purchases among US consumers: a randomized controlled trial. PLoS Med. 2023;20(9):e1004284. pmid:37721952
- 74. Vargas-Meza J, Jáuregui A, Contreras-Manzano A, Nieto C, Barquera S. Acceptability and understanding of front-of-pack nutritional labels: an experimental study in Mexican consumers. BMC Public Health. 2019;19(1):1751. pmid:31888575
- 75. Rahamat S, Jeong E, Arendt SW, Xu Y. Menu labeling influence on purchase behaviors: applying the theory of planned behavior and health consciousness. Appetite. 2022;172:105967. pmid:35157999
- 76. Hearty AP, McCarthy SN, Kearney JM, Gibney MJ. Relationship between attitudes towards healthy eating and dietary behaviour, lifestyle and demographic factors in a representative sample of Irish adults. Appetite. 2007;48(1):1–11. pmid:17049407
- 77. Haws KL, Winterich KP, Naylor RW. Seeing the world through green-tinted glasses: green consumption values and responses to environmentally friendly products. J Consum Psychol. 2014;24:336–54.
- 78. Holm S. A simple sequentially rejective multiple test procedure. Scand J Stat. 1979;6:65–70.
- 79. Wilson DB. Practical Meta Analysis Effect Size Calculator – Regression coefficient from an HLM/Mixed Effects Model (Clustered Data). In: Campbell Collaboration Practical Meta-Analysis Effect Size Calculator [Internet]. 27 Nov 2023 [cited 9 Mar 2026]. Available from: https://www.campbellcollaboration.org/calculator/
- 80.
Cohen J. Statistical power analysis for the behavioral sciences. New York, NY: Academic Press; 2013.
- 81. Ellison B, Lusk JL, Davis D. Looking at the label and beyond: the effects of calorie labels, health consciousness, and demographics on caloric intake in restaurants. Int J Behav Nutr Phys Act. 2013;10:21. pmid:23394433
- 82. Jansen L, van Kleef E, Van Loo EJ. The use of food swaps to encourage healthier online food choices: a randomized controlled trial. Int J Behav Nutr Phys Act. 2021;18(1):156. pmid:34863208
- 83. Schruff-Lim E-M, Van Loo EJ, van der Lans IA, van Trijp HCM. Impact of food swap recommendations on dietary choices in an online supermarket: a randomized controlled trial. Appetite. 2024;194:107158. pmid:38113984
- 84.
Bobo JA. Why smart people make bad food choices: The invisible influences that guide our thinking. Coral Gables, FL: Mango Media; 2022.
- 85.
Robinson TN. Chapter 25 - Stealth interventions for obesity prevention and control: motivating behavior change. In: Dubé L, Bechara A, Dagher A, Drewnowski A, Lebel J, James P, et al, editors. Obesity prevention. San Diego: Academic Press; 2010. pp. 319–327. 10.1016/B978-0-12-374387-9.00025-8
- 86. De Bauw M, Matthys C, Poppe V, Franssens S, Vranken L. A combined Nutri-Score and ‘Eco-Score’ approach for more nutritious and more environmentally friendly food choices? Evidence from a consumer experiment in Belgium. Food Qual Prefer. 2021;93:104276.
- 87. Osman M, Thornton K. Traffic light labelling of meals to promote sustainable consumption and healthy eating. Appetite. 2019;138:60–71. pmid:30880087
- 88. Wolfson JA, Musicus AA, Leung CW, Gearhardt AN, Falbe J. Effect of climate change impact menu labels on fast food ordering choices among US adults: a randomized clinical trial. JAMA Netw Open. 2022;5(12):e2248320. pmid:36574248
- 89. Grummon AH, Zeitlin AB, Lee CJY, Collis C, Cleveland LP, Musicus AA, et al. Ecolabels and the healthfulness and carbon footprint of restaurant meal selections: a randomized clinical trial. JAMA Netw Open. 2025;8(8):e2524773. pmid:40758354
- 90. Parada H Jr, Hall MG, Boynton MH, Brewer NT. Trajectories of responses to pictorial cigarette pack warnings. Nicotine Tob Res. 2017;20(7):876–81.
- 91. Swayampakala K, Thrasher JF, Yong H-H, Nagelhout GE, Li L, Borland R, et al. Over-time impacts of pictorial health warning labels and their differences across smoker subgroups: results from adult smokers in Canada and Australia. Nicotine Tob Res. 2017;20(7):888–96.
- 92. Hitchman SC, Driezen P, Logel C, Hammond D, Fong GT. Changes in effectiveness of cigarette health warnings over time in Canada and the United States, 2002–2011. Nicotine Tob Res. 2014;16:536–43.
- 93. Allcott H, Kessler JB. The welfare effects of nudges: a case study of energy use social comparisons. Am Econ J Appl Econ. 2019;11:236–76.
- 94. Thunström L. Welfare effects of nudges: the emotional tax of calorie menu labeling. Judgm decis mak. 2019;14(1):11–25.
- 95.
Peters E. Affect and emotion. In: Fischhoff B, Brewer NT, Downs JS, editors. Communicating risks and benefits: an evidence-based user’s guide. Silver Spring, MD: United States Department of Health and Human Services Food and Drug Administration; 2011.
- 96. Peters E, Lipkus I, Diefenbach MA. The functions of affect in health communications and in the construction of health preferences. J Commun. 2006;56:S140–62.
- 97. Slovic P, Finucane ML, Peters E, MacGregor DG. Risk as analysis and risk as feelings: some thoughts about affect, reason, risk, and rationality. Risk Anal. 2004;24(2):311–22. pmid:15078302
- 98. Coppock A, Leeper TJ, Mullinix KJ. Generalizability of heterogeneous treatment effect estimates across samples. Proc Natl Acad Sci U S A. 2018;115(49):12441–6. pmid:30446611
- 99. Jeong M, Zhang D, Morgan JC, Ross JC, Osman A, Boynton MH, et al. Similarities and differences in tobacco control research findings from convenience and probability samples. Ann Behav Med. 2019;53(5):476–85. pmid:30052702
- 100. Coppock A. Generalizing from survey experiments conducted on mechanical turk: a replication approach. PSRM. 2018;7(3):613–28.
- 101. Weinberg JD, Freese J, McElhattan D. Comparing data characteristics and results of an online factorial survey between a population-based and a crowdsource-recruited sample. Sociol Sci. 2014;1:292–310.
- 102. Krupnikov Y, Levine AS. Cross-sample comparisons and external validity. J Exp Polit Sci. 2014;1:59–80.