Abstract
Importance:
Despite recent growth in online redemption of Supplemental Nutrition Assistance Program (SNAP) benefits, no previous work has tested the impact of economic and behavioral economic strategies on food purchasing behaviors in an online grocery retail setting among adults with low income.
Objective:
To examine the extent to which financial incentives and default shopping cart options influence fruit and vegetable purchases.
Design:
Randomized controlled experiment.
Setting:
Experimental online grocery store.
Participants:
Adults who currently or have ever received SNAP benefits.
Intervention:
Random assignment to one of four conditions: no intervention, 50% discount on eligible fruits and vegetables, pre-filled shopping carts with tailored fruit and vegetable items (i.e., default options), or a combination of a discount and default options. From October-November 2021, participants were instructed to shop for a week’s worth of groceries for their household, with a budget tailored to household size; no payment was taken.
Results:
Average age of participants (n=2,744) was 46.7 (SD=16.0) years and 53% identified as female. Approximately 67% reported currently receiving SNAP benefits and 54% reported shopping online for groceries in the previous 12 months. On average, participants spent 20.5% (SD=23.5) of total dollars on eligible fruits and vegetables. Compared to no intervention, those in the discount, default, and combination conditions spent 4.7% (98.3% CI: 1.7, 7.7; p<0.001), 7.8% (98.3% CI: 4.8, 10.7; p<0.001), and 13.0% (98.3% CI: 10.0, 16.0; p<0.001) more of total dollars on eligible fruits and vegetables, respectively. There was no difference between the discount and the default conditions (p=0.06) but the effect in the combination condition was significantly larger than both discount and default conditions (p’s<0.001). Default shopping cart items were purchased by 93% and 96% of participants in the default and combination conditions, respectively; whereas, 46% and 53% of participants in the control and discount conditions, respectively, purchased those items (p’s<0.001). We observed no significant variation by age, gender, or race/ethnicity, and results were similar when we excluded those who reported never shopping online for groceries.
Conclusions and Relevance:
In this randomized trial, financial incentives for fruits and vegetables and default options, especially in combination, lead to meaningful increases in online fruit and vegetable purchases among adults with low income.
INTRODUCTION
Adherence to healthy eating patterns plays a critical role in mitigating cardiovascular disease risk.1,2 Adults with low income, including Supplemental Nutrition Assistance Program (SNAP) participants,3 have poorer diet behaviors than higher income adults.4 Thus, interventions designed to increase the affordability of healthy foods, beyond the provision of nutrition assistance benefits, and other non-economic strategies, may help improve diet behaviors in SNAP households. Two popular and effective approaches for increasing healthy food purchases are financial incentives and behavioral nudges.5
Nutrition incentive programs that match SNAP dollars have been shown to increase sales of produce in brick-and-mortar supermarkets,6–8 farmers’ markets,9 and mobile produce markets.10,11 A smaller body of evidence suggests that such incentives also lead to modest improvements in fruit and vegetable consumption,12,13. This body of evidence has supported the introduction and passage of legislation to scale incentives, including the Rhode Island Public Assistance Act.14 Previous studies, however, often do not use an unbiased study design, or do not collect data from a large, national sample. Importantly, no previous work has tested the impact of a fruit and vegetable incentive in an online grocery retail setting.
Interventions informed by behavioral economics may similarly increase healthy food purchasing by “nudging” consumers in low-cost ways that preserve their freedom of choice.15 One powerful insight from behavioral economics is that people tend to stick with default options. This bias can be leveraged to promote health by changing environmental defaults to healthy choices that people can opt out of, such as legislative requirements for restaurants to serve healthy drinks as the default with kid’s meals.16 Default options have been shown to increase healthy food orders in a full-service restaurant,17 as well as healthy food purchases in laboratory-based and online supermarket settings.18–20 Most previous research has had small sample sizes, or has not tailored default options to past purchasing decisions. And no research has focused on pre-filling online shopping carts with fruit and vegetable items.
Studying interventions in an online context is especially important given the recent surge in online grocery sales – which represented 10% of all U.S. grocery sales in 2021, triple the amount from 201721 – and the rapid expansion of the SNAP Online Purchasing Pilot, a program that authorizes retailers to accept SNAP benefits in online transactions.22 Though shoppers with low income are less likely to shop online than their higher-income counterparts,23,24 online redemption of SNAP benefits increased from 0.1% to 3% of total sales from just February 2020 to December 2020.25 As opportunities to use nutrition assistance benefits in online transactions continue to grow, it is critical to test interventions designed to promote healthy online food purchases.
The goal of this study was to examine the extent to which a discount for fruits and vegetables and a default shopping cart strategy, separately and combined, would change expenditures on fruits and vegetables in an online randomized controlled experiment. We hypothesized that non-discounted dollars spent on fruits and vegetables would be higher among participants exposed to the discount plus default options than the control group, and the effect would be larger than other experimental conditions. We also explored whether the impact differed by age group, gender, race/ethnicity, and socioeconomic status.
METHODS
Sample
We used a survey research firm, CloudResearch, to recruit eligible participants from a convenience sample. This approach can yield generalizable findings for online experiments, and studies have shown that data quality is similar to studies conducted via probability-based sampling.26–28 CloudResearch provides participants with points that can be redeemed for various cash and non-cash incentives, and employs quality control mechanisms to minimize fraudulent responses. We instructed CloudResearch to recruit a national sample of adults aged ≥18 years who have ever received SNAP benefits, read and speak English, and live with fewer than four people (to maximize incentives). The sample was recruited to approximately match the distribution of gender and age of U.S. adults in 2019.29 As indicated in our pre-registration, we excluded participants who reused the same IP address, finished the survey in under one-third of median completion time, and/or did not finish the survey or shopping task.
Procedures
We used Qualtrics, an online survey platform,30 to create and distribute a survey, which was completed on a personal computer, laptop, tablet, or mobile phone in October-November 2021. The survey included questions about sociodemographic characteristics, food security, fruit and vegetable intake, and diet and food shopping behaviors(Supplemental File 1); details described elsewhere.31 After completing the survey, participants were randomized to one of four conditions and then asked to click the link to our experimental online grocery store (named ‘Lola’s Grocery’). A detailed description of methods for acquiring and cleaning online store data and the design of our two-dimensional platform, and its validity and acceptability, are published elsewhere.32 Briefly, we used an platform called Gorilla developed by Cauldron Science, Ltd.,33 which mimics the appearance and functionalities of a top U.S. online grocery retailer, including browsing, search, product pages, shopping cart, and checkout. The store has over 20,000 products, organized by department, aisle, and shelf, with product images, price, and nutrition information.
Prior to shopping, participants were instructed to select a typical week’s worth of food for their household, with details about the budget (no payment was taken) and, if relevant, additional instructions related to the discounts and/or default shopping cart items for participants randomized to the experimental conditions (Supplemental File 2). Their shopping budget was based on food-at-home expenditure data from the National Household Food Acquisition and Purchase Survey (FoodAPS), a nationally-representative survey of food purchases of SNAP and non-SNAP households in 2012–2013.34 After accounting for inflation,35 we estimated that SNAP households with one, two, three, or four total persons spend on average $60, $100, $120, or $130 per week on groceries, respectively. After completing the shopping task, participants answered eight questions related to the process of shopping in our store,36 and the degree to which they would support default options as a retailer strategy, with or without discounts.
To incentivize participation and truthful responses, participants were notified they would be automatically entered into a lottery upon completion of the shopping task and that 100 participants would have their cart items delivered to their household. At the end of the study, however, we revealed to all participants that individuals who won the lottery were instead provided a gift card with the equivalent amount of money loaded onto it that they spent in the study. This approach was used to mitigate potential issues related to delivery, and to minimize the collection of personally-identifiable information.
Experimental conditions
All participants were randomly assigned to one of four conditions: no intervention, 50% discount on eligible fruits and vegetables, shopping carts pre-filled with tailored fruit and vegetable items (i.e., default options), or a combination of a discount and default options. To mimic existing incentive programs,37,38 participants randomized to the discount condition received half-off qualifying fruits and vegetables, up to a value of $20 off. Fruits and vegetables that qualified for the discount were labeled as eligible in the store, and the discount was applied to items at the point-of-selection. Based on healthy eating guidelines,39 eligible items included fresh, frozen, and canned fruits and vegetables with ≤140 milligrams of sodium per serving, and prepared produce intended for off-premises consumption. Ineligible items included herbs and spices, dried fruits and vegetables, fruit and vegetable juices, and fruits and vegetables with added sugars or fat.
Participants randomized to the default condition had their cart pre-filled with one fruit product and one vegetable product, which they could remove at any point while shopping. Default items were based on expenditure data from FoodAPS, which we used to identify the five most frequently purchased fruits (apples, bananas, grapes oranges, strawberries) and vegetables (carrots, green beans, lettuce, onions, tomatoes) in SNAP-participating households. We then identified 10 corresponding products from our online store database. We chose products with the highest “shelf rank,” an indicator of a product’s order on a shelf when sorted by best seller. In the survey, participants were asked to select which of the five fruits and five vegetables they purchased the most frequently, and we used their selections to tailor their default options, for a total of 25 possible combinations of default products (Supplemental File 3). For participants randomized to receive both a discount and default options, the pre-filled products were discounted by 50%. Participants randomized to the control condition received no discount or default products.
Outcomes
To account for differences in item prices across conditions and shopping budget by household size, our primary outcome was the percentage of non-discounted dollars spent on eligible fruit and vegetables per basket (i.e., dollars spent before subtracting a discount). Secondary outcomes included non-discounted and out-of-pocket dollars spent on fruits and vegetables; non-discounted dollars spent on fruits only, vegetables only, fresh fruits and vegetables, frozen fruits and vegetables, and canned fruits and vegetables; and total calories, fat (g), saturated fat (g), carbohydrates (g), fiber (g), protein (g), and salt (mg) from fruits and vegetables. We calculated spending (or nutrient value) in each category by multiplying the price (or nutrient value) per item by the quantity purchased of that item, then summed across all items purchased in that category.
Statistical analyses
We used linear regression to regress the outcome variable on indicator variables for each of the experimental conditions, with the control condition as referent. Analyses indicated that enrollment of 1,264 adults would provide 90% power to detect a 9% difference between any experimental condition and the control condition, including sufficient power for a two-stage approach to hypothesis testing. This allowed us to first assess the effect of experimental conditions compared to control at a Type I error rate of 0.017, and then compare successful experimental conditions to each other using a Tukey’s Honest Significant Difference approach; this allowed us to detect a 2% difference between experimental conditions. Pre-specified analyses indicated that a sample size of 2,528 adults also allowed for examination of interaction effects by age, gender, and race/ethnicity. Race/ethnicity was self-reported by participants using National Health and Nutrition Examination Survey categories (American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or Pacific Islander, White, Other). Given differences in household shopping responsibilities and internet use by gender, age, and race/ethnicity,40 we hypothesized that effects would be larger among female, younger, and White adults. In post-hoc analyses, we also examined interaction effects by income, education, food security status.
To assess robustness of our results to differences in online grocery shopping history, we conducted a sensitivity analysis excluding participants who reported they had never shopped online for groceries. We also used logistic regression to assess differences in the likelihood of purchasing a default shopping cart item (from the shopping cart or store aisles) by condition. We calculated the percentage of participants who endorsed positive statements about their experience in the store. To assess attrition bias, we compared the characteristics of those who completed the survey task but not the shopping task to those who completed both tasks. All analyses were conducted using Stata 16.0. This study was approved by the NYU Grossman School of Medicine Institutional Review Board and participants provided electronic informed consent. Reporting followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guidelines.
RESULTS
We excluded participants who reused the same IP address (n=265), finished the pre-shopping survey in under one-third of median completion time of 6.4 minutes (i.e., <2.1 minutes) (n=51), did not finish the survey (n=90), and/or did not finish the shopping task (n=782) (Figure 1). The final sample included 2,744 adults who had ever received SNAP benefits. Average age of participants was 46.7 (SD=16.0) years and 53% identified as female (Table 1; Supplemental Table 1). Median household size was 2 (IQR=1, 3) persons, including 1 (IQR=1, 2) child below 18 years. Approximately 67% of participants reported currently receiving SNAP benefits and 70% were classified as food insecure. A little over half of participants (54%) reported shopping online for groceries (Supplemental Table 2). Participants’ characteristics were balanced across conditions. Those who completed the survey and shopping task (versus survey only) were more likely to be female (53.5% vs. 45.2%; p<0.001) and have at least some college education (60.1% vs. 51.7%; p<0.001).
TABLE 1.
All (n=2744) | Control (n=649) | Default (n=727) | Discount (n=682) | Default + Discount (n=686) | |
---|---|---|---|---|---|
Mean (SD) or n (%) | Mean (SD) or n (%) | Mean (SD) or n (%) | Mean (SD) or n (%) | Mean (SD) or n (%) | |
Age | 46.7 (16.0) | 45.9 (15.4) | 47.0 (16.3) | 47.2 (16.1) | 46.7 (16.1) |
Female | 1447 (52.7%) | 344 (53.0%) | 396 (54.5%) | 346 (50.7%) | 361 (52.6%) |
Household size, total | 2.3 (1.0) | 2.3 (1.0) | 2.3 (1.0) | 2.2 (1.0) | 1.1 (2.3) |
Household size, children <18 years | 1.4 (0.7) | 1.4 (0.7) | 1.4 (0.7) | 1.4 (0.7) | 0.7 (1.4) |
Hispanic, Latino, or Spanish | 269 (9.8%) | 63 (9.7%) | 75 (10.3%) | 67 (9.8%) | 64 (9.3%) |
Race | |||||
American Indian or Alaska Native | 39 (1.4%) | 9 (1.4%) | 6 (0.8%) | 18 (2.6%) | 6 (0.9%) |
Asian | 49 (1.8%) | 13 (2.0%) | 10 (1.4%) | 10 (1.5%) | 16 (2.3%) |
Black or African American | 390 (14.2%) | 99 (15.3%) | 93 (12.8%) | 91 (13.3%) | 107 (15.6%) |
Native Hawaiian or Pacific Islander | 8 (0.3%) | 2 (0.3%) | 2 (0.3%) | 516 (75.7%) | 4 (0.6%) |
White | 2074 (75.6%) | 490 (75.5%) | 562 (77.3%) | 18 (2.6%) | 506 (73.8%) |
Other | 70 (2.6%) | 20 (3.1%) | 15 (2.1%) | 1 (0.1%) | 17 (2.5%) |
More than 1 | 12 (0.4%) | 16 (2.5%) | 4 (0.6%) | 28 (4.1%) | 7 (1.0%) |
Prefer not to answer | 102 (3.7%) | - | 35 (4.8%) | - | 23 (3.4%) |
Education | |||||
Less than 9th grade | 17 (0.6%) | 4 (0.6%) | 4 (0.6%) | 3 (0.4%) | 6 (0.9%) |
9th to 12th grade - No diploma | 150 (5.5%) | 41 (6.3%) | 36 (5.0%) | 38 (5.6%) | 35 (5.1%) |
High school graduate | 749 (27.3%) | 180 (27.7%) | 194 (26.7%) | 175 (25.7%) | 200 (29.2%) |
GED or equivalent | 182 (6.6%) | 42 (6.5%) | 53 (7.3%) | 43 (6.3%) | 44 (6.4%) |
Some college, no degree | 859 (31.3%) | 201 (31.0%) | 227 (31.2%) | 216 (31.7%) | 215 (31.3%) |
Associate’s degree | 367 (13.4%) | 83 (12.8%) | 107 (14.7%) | 97 (14.2%) | 80 (11.7%) |
Bachelor’s degree | 307 (11.2%) | 69 (10.6%) | 80 (11.0%) | 80 (11.7%) | 78 (11.4%) |
Graduate or professional degree | 110 (4.0%) | 28 (4.3%) | 26 (3.6%) | 29 (4.3%) | 27 (3.9%) |
Prefer not to answer | 3 (0.1%) | 1 (0.2%) | - | 1 (0.1%) | 1 (0.1%) |
Income | |||||
<$20,000 | 1187 (43.3%) | 266 (41.0%) | 307 (42.2%) | 297 (43.5%) | 317 (46.2%) |
$20,000–39,999 | 1017 (37.1%) | 247 (38.1%) | 278 (38.2%) | 257 (37.7%) | 235 (34.3%) |
$40,000–59,999 | 348 (12.7%) | 84 (12.9%) | 93 (12.8%) | 90 (13.2%) | 81 (11.8%) |
$60,000-$79,999 | 108 (3.9%) | 25 (3.9%) | 30 (4.1%) | 25 (3.7%) | 28 (4.1%) |
$80,000-$99,999 | 32 (1.2%) | 11 (1.7%) | 7 (1.0%) | 4 (0.6%) | 10 (1.5%) |
$100,000–119,999 | 10 (0.4%) | 2 (0.3%) | 2 (0.3%) | 1 (0.1%) | 5 (0.7%) |
$120,000 to $139,999 | 3 (0.1%) | 2 (0.3%) | 1 (0.1%) | 1 (0.1%) | 2 (0.3%) |
$140,000 to $159,999 | 4 (0.1%) | 3 (0.5%) | 3 (0.4%) | 1 (0.1%) | 2 (0.3%) |
$160,000 to $179,999 | 3 (0.1%) | 2 (0.3%) | 5 (0.7%) | 1 (0.1%) | 3 (0.4%) |
$180,000 to $199,999 | 3 (0.1%) | 3 (0.5%) | 1 (0.1%) | 1 (0.1%) | 3 (0.4%) |
≥$200,000 | 6 (0.2%) | 4 (0.6%) | - | 1 (0.1%) | - |
Don’t know | 12 (0.4%) | - | - | 3 (0.4%) | - |
Prefer not to answer | 11 (0.4%) | - | - | - | - |
Marital status | |||||
Married | 618 (22.5%) | 163 (25.1%) | 141 (19.4%) | 155 (22.7%) | 159 (23.2%) |
Widowed | 191 (7.0%) | 47 (7.2%) | 49 (6.7%) | 44 (6.5%) | 51 (7.4%) |
Divorced | 476 (17.3%) | 105 (16.2%) | 139 (19.1%) | 120 (17.6%) | 112 (16.3%) |
Separated | 107 (3.9%) | 30 (4.6%) | 23 (3.2%) | 21 (3.1%) | 33 (4.8%) |
Never Married | 920 (33.5%) | 207 (31.9%) | 243 (33.4%) | 241 (35.3%) | 229 (33.4%) |
Living with Partner | 418 (15.2%) | 94 (14.5%) | 131 (18.0%) | 99 (14.5%) | 94 (13.7%) |
Prefer not to answer | 14 (0.5%) | 3 (0.5%) | 1 (0.1%) | 2 (0.3%) | 8 (1.2%) |
Employment | |||||
Working at a job or business | 845 (30.8%) | 193 (29.7%) | 222 (30.5%) | 223 (32.7%) | 207 (30.2%) |
With a job or business but not at work | 70 (2.6%) | 18 (2.8%) | 24 (3.3%) | 11 (1.6%) | 17 (2.5%) |
Looking for work | 501 (18.3%) | 114 (17.6%) | 137 (18.8%) | 133 (19.5%) | 117 (17.1%) |
Not working at a job or business | 1131 (41.2%) | 268 (41.3%) | 296 (40.7%) | 281 (41.2%) | 286 (41.7%) |
Part-time or full-time student | 103 (3.8%) | 30 (4.6%) | 23 (3.2%) | 22 (3.2%) | 28 (4.1%) |
Prefer not to answer | 94 (3.4%) | 26 (4.0%) | 25 (3.4%) | 12 (1.8%) | 31 (4.5%) |
Food insecurityb | 1917 (69.9%) | 449 (69.2%) | 483 (70.8%) | 502 (69.1%) | 483 (70.4%) |
SNAP participation, currently | 1842 (67.1%) | 424 (65.3%) | 494 (68.0%) | 445 (65.2%) | 479 (69.8%) |
Fruit, purchased most oftenc | |||||
Apples | 472 (17.2%) | 102 (15.7%) | 115 (15.8%) | 137 (20.1%) | 118 (17.2%) |
Bananas | 1319 (48.1%) | 310 (47.8%) | 348 (47.9%) | 325 (47.7%) | 336 (49.0%) |
Grapes | 390 (14.2%) | 96 (14.8%) | 114 (15.7%) | 82 (12.0%) | 98 (14.3%) |
Oranges | 232 (8.5%) | 60 (9.2%) | 54 (7.4%) | 60 (8.8%) | 58 (8.5%) |
Strawberries | 331 (12.1%) | 81 (12.5%) | 96 (13.2%) | 78 (11.4%) | 76 (11.1%) |
Vegetable, purchased most oftenc | |||||
Carrots | 409 (14.9%) | 94 (14.5%) | 107 (14.7%) | 96 (14.1%) | 112 (16.3%) |
Lettuce | 726 (26.5%) | 154 (23.7%) | 200 (27.5%) | 186 (27.3%) | 186 (27.1%) |
Onions | 685 (25.0%) | 176 (27.1%) | 190 (26.1%) | 167 (24.5%) | 152 (22.2%) |
String beans | 281 (10.2%) | 70 (10.8%) | 61 (8.4%) | 78 (11.4%) | 72 (10.5%) |
Tomatoes | 643 (23.4%) | 155 (23.9%) | 169 (23.2%) | 155 (22.7%) | 164 (23.9%) |
BRFSS 2017 screener, times per week | |||||
Fruit juice | 2.6 (4.2) | 2.4 (4.0) | 2.5 (4.0) | 2.7 (4.2) | 2.9 (4.6) |
Fruit | 3.9 (4.6) | 3.7 (4.6) | 3.9 (4.7) | 3.8 (4.5) | 4.1 (4.8) |
Beans | 2.0 (3.1) | 2.1 (3.3) | 1.9 (2.9) | 1.9 (3.0) | 2.1 (3.3) |
Dark green vegetables | 2.9 (3.6) | 2.9 (3.9) | 2.8 (3.4) | 2.7 (3.3) | 3.1 (3.7) |
Orange-colored vegetables | 2.1 (3.3) | 2.0 (3.3) | 2.0 (3.2) | 2.0 (3.3) | 2.4 (3.6) |
Other vegetables | 3.5 (3.9) | 3.6 (4.1) | 3.4 (3.9) | 3.6 (4.1) | 3.3 (3.5) |
Total | 16.9 (16.8) | 16.8 (17.7) | 16.4 (15.7) | 16.7 (16.4) | 17.9 (17.5) |
Responsible for most of household food shopping | |||||
Yes | 2362 (86.1%) | 555 (85.5%) | 618 (85.0%) | 582 (85.3%) | 607 (88.5%) |
No | 198 (7.2%) | 54 (8.3%) | 52 (7.2%) | 59 (8.7%) | 33 (4.8%) |
No one person is responsible | 184 (6.7%) | 40 (6.2%) | 57 (7.8%) | 41 (6.0%) | 46 (6.7%) |
Responsible for most of household food preparation | |||||
Yes | 2239 (81.6%) | 517 (79.7%) | 590 (81.2%) | 563 (82.6%) | 569 (82.9%) |
No | 327 (11.9%) | 91 (14.0%) | 75 (10.3%) | 86 (12.6%) | 75 (10.9%) |
No one person is responsible | 178 (6.5%) | 41 (6.3%) | 62 (8.5%) | 33 (4.8%) | 42 (6.1%) |
Note: SNAP=Supplemental Nutrition Assistance Program; BRFFS=Behavioral Risk Factor Surveillance System
Sociodemographic and food insecurity questions were derived from the 2017–2018 National Health and Nutrition Examination Survey. We captured weekly fruit and vegetable intake using a 6-item fruit and vegetable dietary intake module from the Behavioral Risk Factor Surveillance System.
Based on the two-item Hunger Vital Sign food security screener, food insecurity defined as yes if a participant indicated that it was true or sometimes true that 1) their household was worried whether their food would run out before they got money to buy more, and/or 2) the food that they bought just didn’t last and they didn’t have enough money to get more.
The five fruit items and five vegetable items were based on expenditure data in SNAP-participating households in the Food Acquisition and Purchase Survey.
About 93% of participants purchased a fruit or vegetable (Table 2) and on average participants spent 20.5% (SD=23.5) of total dollars on fruits and vegetables. Compared to the control condition, those in the discount, default, and combination conditions spent 4.7% (98.3% CI: 1.7, 7.7), 7.8% (98.3% CI: 4.8, 10.7), and 13.0% (98.3% CI: 10.0, 16.0) more on fruits and vegetables, respectively (p’s<0.001) (Table 3). There was no difference between discount and default conditions (p=0.06), but the effect in the combination condition was larger than the other experimental conditions (p’s<0.001). Default cart items were purchased by 93% and 96% of participants in the default and combination conditions, respectively, compared to 46% and 53% of participants in the control and discount conditions, respectively (p’s<0.001).
TABLE 2.
Mean (SD) or n (%) | Mean (SD) or n (%) | Mean (SD) or n (%) | Mean (SD) or n (%) | Mean (SD) or n (%) | |
---|---|---|---|---|---|
| |||||
FULL SAMPLE | All (n=2744) | Control (n=649) | Default (n=727) | Discount (n=682) | Default + Discount (n=686) |
Primary outcome | |||||
Percentage of non-discounted dollars spent on eligible fruits and vegetables, all | 20.5% (23.5) | 14.0% (16.2) | 21.8% (25.5) | 18.7% (18.8) | 27.1% (28.7) |
Secondary outcomes | |||||
Non-discounted dollars spent on eligible fruits and vegetables, all | $13.18 (13.79) | $11.31 (12.82) | $11.68 (11.36) | $15.27 (16.08) | $14.47 (14.18) |
Out-of-pocket dollars spent on eligible fruits and vegetables, all | $9.73 (11.18) | $11.31 (12.82) | $11.68 (11.36) | $8.25 (10.78) | $7.66 (8.96) |
Non-discounted dollars spent on eligible fruits | $5.45 (8.08) | $4.28 (7.14) | $4.85 (6.78) | $6.51 (9.94) | $6.12 (7.91) |
Non-discounted dollars spent on eligible vegetables | $7.74 (9.64) | $7.03 (9.34) | $6.84 (7.93) | $8.76 (11.04) | $8.34 (9.94) |
Non-discounted dollars spent on eligible fresh fruits and vegetables | $11.39 (12.52) | $9.61 (11.83) | $10.49 (10.67) | $12.94 (14.56) | $12.48 (12.51) |
Non-discounted dollars spent on eligible frozen fruits and vegetables | $1.60 (4.63) | $1.48 (4.01) | $1.05 (3.23) | $2.07 (5.49) | $1.83 (5.38) |
Non-discounted dollars spent on eligible canned fruits and vegetables | $0.19 (0.90) | $0.22 (0.92) | $0.13 (0.65) | $0.27 (1.20) | $0.16 (0.75) |
Total calories from fruits and vegetables, kcal | 1505.5 (1574.8) | 1298.1 (1595.0) | 1496.8 (1472.8) | 1526.2 (1641.3) | 1690.5 (1572.0) |
Total fat from fruits and vegetables, g | 9.1 (21.8) | 7.6 (21.1) | 8.3 (20.4) | 10.0 (22.8) | 10.3 (22.6) |
Total saturated fat from fruits and vegetables, g | 1.2 (3.0) | 1.0 (2.8) | 1.1 (2.7) | 1.2 (3.1) | 1.4 (3.3) |
Total carbohydrates from fruits and vegetables, g | 346.4 (359.5) | 294.4 (367.9) | 349.9 (335.6) | 345.9 (372.7) | 392.3 (357.0) |
Total salt from fruits and vegetables, mg | 414.5 (680.4) | 380.5 (656.1) | 376.7 (616.2) | 451.0 (746.7) | 450.4 (696.0) |
Total fiber from fruits and vegetables, g | 57.5 (54.9) | 47.9 (53.1) | 56.1 (48.6) | 58.9 (58.8) | 66.5 (57.4) |
Total protein from fruits and vegetables, g | 36.8 (42.8) | 34.2 (43.4) | 35.2 (40.6) | 38.3 (44.3) | 39.3 (42.7) |
Participants who purchased fruits and/or vegetables, % | 2554 (93.1%) | 550 (84.7%) | 714 (98.2%) | 612 (89.7%) | 678 (98.8%) |
Participants who purchased default shopping cart items, % | 1992 (72.6%) | 297 (45.8%) | 679 (93.4%) | 361 (52.9%) | 655 (95.5%) |
| |||||
FEMALE PARTICIPANTS | All (n=1447) | Controls (n=344) | Defaults (n=396) | Discounts (n=361) | Defaults + Discounts (n=346) |
| |||||
Primary outcome | |||||
Percentage of non-discounted dollars spent on eligible fruits and vegetables, all | 20.8% (25.3%) | 14.6% (18.6%) | 23.0% (27.6%) | 17.4% (19.3%) | 27.8% (31.1%) |
Secondary outcomes | |||||
Non-discounted dollars spent on eligible fruits and vegetables, all | $14.25 (13.46) | $12.05 (12.31) | $12.10 (10.63) | $17.00 (14.99) | $16.08 (14.95) |
Out-of-pocket dollars spent on eligible fruits and vegetables, all | $10.48 (10.67) | $12.05 (12.31) | $12.10 (10.63) | $9.03 (9.18) | $8.61 (9.86) |
Non-discounted dollars spent on eligible fruits | $5.66 (7.11) | $4.21 (5.80) | $4.97 (5.17) | $7.02 (8.19) | $6.50 (8.52) |
Non-discounted dollars spent on eligible vegetables | $8.59 (9.72) | $7.84 (9.36) | $7.12 (7.80) | $9.98 (11.00) | $9.57 (10.36) |
Non-discounted dollars spent on eligible fresh fruits and vegetables | $12.27 (12.30) | $10.07 (11.59) | $10.75 (9.82) | $14.60 (13.75) | $13.80 (13.33) |
Non-discounted dollars spent on eligible frozen fruits and vegetables | $1.73 (4.53) | $1.66 (3.72) | $1.15 (3.34) | $2.06 (4.97) | $2.10 (5.71) |
Non-discounted dollars spent on eligible canned fruits and vegetables | $0.25 (1.05) | $0.32 (1.16) | $0.19 (0.79) | $0.34 (1.42) | $0.17 (0.70) |
Total calories from fruits and vegetables, kcal | 1653.0 (1642.0) | 1454.3 (1727.3) | 1561.8 (1446.9) | 1721.9 (1726.0) | 1876.1 (1653.3) |
Total fat from fruits and vegetables, g | 10.5 (23.5) | 8.6 (23.3) | 9.5 (20.0) | 13.0 (26.4) | 11.2 (24.2) |
Total saturated fat from fruits and vegetables, g | 1.4 (3.2) | 1.1 (3.0) | 1.3 (2.7) | 1.6 (3.6) | 1.5 (3.5) |
Total carbohydrates from fruits and vegetables, g | 378.3 (375.9) | 327.5 (397.1) | 363.8 (334.4) | 387.8 (390.4) | 433.3 (377.8) |
Total salt from fruits and vegetables, mg | 489.3 (754.7) | 444.5 (728.7) | 411.0 (653.5) | 570.5 (861.9) | 540.1 (764.0) |
Total fiber from fruits and vegetables, g | 63.8 (57.9) | 53.7 (58.0) | 58.8 (47.8) | 68.8 (62.7) | 74.0 (61.2) |
Total protein from fruits and vegetables, g | 41.2 (44.4) | 39.5 (47.0) | 37.0 (39.8) | 43.3 (46.0) | 45.5 (44.8) |
Participants who purchased fruits and/or vegetables, % | 1374 (95.0%) | 301 (87.5%) | 391 (98.7%) | 337 (93.4%) | 344 (99.4%) |
Participants who purchased default shopping cart items, % | 1085 (75.0%) | 165 (48.0%) | 371 (93.7%) | 211 (58.4%) | 333 (96.1%) |
| |||||
MALE PARTICIPANTS | All (n=1286) | Controls (n=304) | Defaults (n=727) | Discounts (n=321) | Defaults + Discounts (n=332) |
| |||||
Primary outcome | |||||
Percentage of non-discounted dollars spent on eligible fruits and vegetables, all | 20.3% (21.7%) | 13.6% (13.7%) | 20.9% (23.7%) | 20.1% (18.3%) | 26.3% (26.3%) |
Secondary outcomes | |||||
Non-discounted dollars spent on eligible fruits and vegetables, all | $12.01 (14.10) | $10.50 (13.35) | $11.24 (12.19) | $13.57 (17.06) | $12.61 (13.06) |
Out-of-pocket dollars spent on eligible fruits and vegetables, all | $8.93 (11.72) | $10.50 (13.35) | $11.24 (12.19) | $7.49 (12.24) | $6.57 (7.73) |
Non-discounted dollars spent on eligible fruits | $5.22 (9.06) | $4.37 (8.43) | $4.72 (8.34) | $6.02 (11.52) | $5.70 (7.20) |
Non-discounted dollars spent on eligible vegetables | $6.79 (9.46) | $6.12 (9.26) | $6.52 (8.09) | $7.55 (11.00) | $6.91 (9.21) |
Non-discounted dollars spent on eligible fresh fruits and vegetables | $10.42 (12.72) | $9.11 (12.12) | $10.24 (11.62) | $11.29 (15.25) | $10.93 (11.37) |
Non-discounted dollars spent on eligible frozen fruits and vegetables | $1.47 (4.75) | $1.29 (4.32) | $0.93 (3.11) | $2.09 (6.01) | $1.53 (5.01) |
Non-discounted dollars spent on eligible canned fruits and vegetables | $0.12 (0.68) | $0.10 (0.50) | $0.07 (0.40) | $0.19 (0.90) | $0.14 (0.79) |
Total calories from fruits and vegetables, kcal | 1342.6 (1478.7) | 1125.2 (1414.5) | 1424.6 (1504.0) | 1333.8 (1530.0) | 1473.5 (1441.3) |
Total fat from fruits and vegetables, g | 7.4 (19.6) | 6.5 (18.4) | 6.9 (20.9) | 6.9 (18.0) | 9.3 (20.8) |
Total saturated fat from fruits and vegetables, g | 1.0 (2.6) | 0.8 (2.5) | 0.9 (2.6) | 0.9 (2.5) | 1.3 (3.0) |
Total carbohydrates from fruits and vegetables, g | 311.2 (336.8) | 257.9 (328.7) | 334.6 (337.1) | 304.7 (349.7) | 344.3 (325.2) |
Total salt from fruits and vegetables, mg | 330.9 (575.3) | 309.2 (556.3) | 337.4 (568.0) | 329.2 (584.1) | 346.5 (593.1) |
Total fiber from fruits and vegetables, g | 50.5 (50.4) | 41.5 (46.1) | 53.1 (49.4) | 49.1 (52.7) | 57.8 (51.6) |
Total protein from fruits and vegetables, g | 31.8 (40.3) | 28.2 (38.3) | 33.2 (41.7) | 33.5 (42.2) | 32.1 (38.7) |
Participants who purchased fruits and/or vegetables, % | 1170 (91.0%) | 248 (81.6%) | 322 (97.9%) | 276 (85.8%) | 326 (98.1%) |
Participants who purchased default shopping cart items, % | 900 (70.0%) | 131 (43.1%) | 307 (93.3%) | 153 (47.6%) | 314 (94.7%) |
TABLE 3.
Control (n=649) | Default (n=727) | Discount (n=682) | Default + Discount (n=686) | |
---|---|---|---|---|
β (98.3% CI) | β (98.3% CI) | β (98.3% CI) | β (98.3% CI) | |
Primary outcome | ||||
Percentage of non-discounted dollars spent on eligible fruits and vegetables, all | - | 7.8 (4.8–10.7) | 4.7 (1.7–7.7) | 13.0 (10.0–16.0) |
Secondary outcomes | ||||
Non-discounted dollars spent on eligible fruits and vegetables, all | - | $0.37 (−1.39–2.14) | $3.96 (2.17–5.76) | $3.16 (1.37–4.95) |
Out-of-pocket dollars spent on eligible fruits and vegetables, all | - | $0.37 (−1.05–1.80) | −$3.06 (−4.51−-1.62) | −$3.65 (−5.10−-2.21) |
Non-discounted dollars spent on eligible fruits | - | $0.57 (−0.47–1.60) | $2.23 (1.18–3.28) | $1.85 (0.80–2.90) |
Non-discounted dollars spent on eligible vegetables | - | −$0.20 (−1.43–1.04) | $1.73 (0.47–2.99) | $1.31 (0.06–2.57) |
Non-discounted dollars spent on eligible fresh fruits and vegetables | - | $0.88 (−0.72–2.49) | $3.33 (1.70–4.96) | $2.87 (1.24–4.50) |
Non-discounted dollars spent on eligible frozen fruits and vegetables | - | −$0.43 (−1.03–0.16) | $0.59 (−0.02–1.19) | $0.34 (−0.26–0.95) |
Non-discounted dollars spent on eligible canned fruits and vegetables | - | −$0.08 (−0.20–0.03) | $0.05 (−0.07–0.17) | −$0.06 (−0.18–0.06) |
Total calories from fruits and vegetables, kcal | - | 198.7 (−3.7–401.1) | 228.1 (22.6–433.7) | 392.4 (187.1–597.7) |
Total fat from fruits and vegetables, g | - | 0.7 (−2.1–3.5) | 2.4 (−0.5–5.2) | 2.7 (−0.1–5.6) |
Total saturated fat from fruits and vegetables, g | - | 0.1 (−0.3–0.5) | 0.3 (−0.1–0.7) | 0.4 (0.0–0.8) |
Total carbohydrates from fruits and vegetables, g | - | 55.5 (9.3–101.7) | 51.5 (4.6–98.4) | 97.9 (51.1–144.7) |
Total salt from fruits and vegetables, mg | - | −3.8 (−91.4–83.9) | 70.5 (−18.5–159.5) | 70.0 (−18.9–158.9) |
Total fiber from fruits and vegetables, g | - | 8.2 (1.1–15.2) | 11.0 (3.9–18.2) | 18.6 (11.4–25.7) |
Total protein from fruits and vegetables, g | - | 1.0 (−4.5–6.5) | 4.2 (−1.4–9.8) | 5.2 (−0.4–10.7) |
Average non-discounted dollars and out-of-pocket dollars spent on eligible fruits and vegetables across conditions was $13.18 (SD=$13.79) and $9.73 (SD=11.18), respectively (Table 2). Spending in non-discounted dollars was higher in the discount condition [$3.16 (98.3% CI: 1.37, 4.95] and combination condition [$3.96 (98.3% CI: 2.17, 5.76], but not the default condition [$0.37 (98.3% CI: −1.39, 2.14)] (Table 3). Out-of-pocket spending was lower among those in the discount condition [-$3.65 (98.3% CI: −5.10, −2.21)] and combination condition [-$3.06 (98.3% CI: −4.51, −1.62)], but not the default condition [$0.37 (98.3% CI: −1.05, 1.80)]. Spending on fruits only, vegetables only, and fresh fruits and vegetables was higher among participants randomized to receive discounts (alone or in combination) versus no intervention. Total calories, carbohydrates, and fiber from fruits and vegetables were significantly higher in the discount and combination conditions compared to no intervention. We did not observe meaningful differences in expenditures on frozen or canned fruits and vegetables across conditions, or other nutrient groups.
We observed no interaction effects by age, gender, race/ethnicity, income, education, or food security status; and no differences when excluding those who reported never shopping online for groceries (Supplemental Table 3). A majority of participants reported that the store felt like a real online grocery store (86%), there were enough food and beverage options (77%), their purchases were similar to their regular purchases (75%), they could imagine doing their real-life shopping in our store (76%), and they could easily find all of the items they were looking for (74%) (Supplemental Table 4). There was a high degree of support for default options as a retailer strategy, with (67%) or without (73%) discounts.
DISCUSSION
Previous research has found that nutrition labels,41–44 product placement,43,45–47 and offering healthy alternatives43,45 promote healthier food choices in online supermarkets. We found that financial incentives are also effective in promoting online fruit and vegetable purchases in a large, diverse sample of individuals with low income. The effect from our discount alone (4.7%) was smaller compared to the Healthy Incentives Pilot, which found an 11% increase in total fruit and vegetable purchases with a 30% discount.7 This may be partially due to higher reluctance to purchase produce online versus in-store.31 Yet, our results are consistent with previous work evaluating financial incentives in other in-person retail settings,8,11 including a 50% discount program in Michigan supermarkets, which resulted in a 2.2% to 7.4% increase in fruit and vegetable purchases over two years. The expansion of existing incentive programs to online settings is a promising strategy to promote equitable access to food, and, given the decrease in participants’ out-of-pocket spending on fruits and vegetables, may increase purchasing power for those with food insecurity – especially in combination with more cost-effective strategies like sugary drink restrictions and economic disincentives.49
We also observed a similar impact on purchases from a behavioral nudge that pre-filled participants’ carts with default fruit and vegetable products based on their shopping history. These default options were successful in getting participants to purchase the default products in their cart, with over 90% of participants in a default condition purchasing a default item, compared to about half of participants in other conditions. Participants also expressed a high degree of support for default options as a retailer strategy. Taken together, these results suggest that default options are effective in motivating individuals to purchase fruits and vegetables.
We also found evidence that combining discounts and default options had a synergistic effect, which provides support for the implementation of multiple healthy eating strategies in an online setting. Given how changes in expenditures translated into positive changes in fruit and vegetable intake in the Healthy Incentives Pilot,50 we expect that effects observed in this study would meaningfully improve population health. Yet, we are not aware of legislation or programs related to default options in the real world, which may reflect a lack of evidence and/or interest from retailers, potentially due to concerns about negative customer perceptions.
Limitations
Participants were recruited online, so they may have been more likely to shop online than the broader SNAP population. However, results did not differ when we excluded those who reported never shopping online for groceries. The sample may not have been representative of the target population, given how non-English speakers were excluded and those who started the shopping task were slightly more likely be female and have higher education. It is possible that hypothetical shopping choices do not reflect actual purchases, but we informed participants they were entered into a lottery to receive the items in their cart, so we expect participants selected items they actually wanted to receive. Furthermore, a large majority reported their purchases were similar to their regular purchases.
Conclusions
Our results support the use of financial incentives as a way to meaningfully increase fruit and vegetable purchases for those shopping with SNAP benefits online. We also found evidence that prefilling shopping carts with fruits and vegetables motivated individuals to purchase default products, with a synergistic effect with discounts. Future research should explore the effectiveness of strategies designed to mitigate lack of trust, and other non-economic barriers, on purchases of fresh products online. Evaluating other interventions in online settings, such as restrictions on targeted marketing, is also a valuable next step.
Supplementary Material
KEY POINTS.
Question:
Do financial incentives and default shopping cart options promote fruit and vegetable purchases among adults with low income shopping in a naturalistic online grocery store?
Findings:
In this randomized trial, participants receiving a 50% discount on eligible fruits and vegetables and/or default fruit and vegetable items in their shopping cart spent significantly more (5–13%) on fruits and vegetables compared to those who did not receive an intervention.
Meaning:
Nutrition incentives programs may be effective in promoting healthy food choices among those using nutrition assistance benefits in online retail settings.
ACKNOWLEDGMENTS:
We would like to thank Jamie Halula, Sophia Hurr, Rhea Naik, Carmen Prestemon, and Carla Seet for their help in cleaning the dataset and reviewing the online grocery store. Additionally, we would like to thank Emily Busey for designing our online grocery store logo, Bridgett Hollingsworth for providing guidance on how to tag red meat and processed red meat, Yiqing Zhang for providing insight into the data scraping process, and Dr. Wilma Waterlander for her general feedback on store development and the process evaluation. We would also like to thank the researchers involved in developing the Woods supermarket platform for their support, including Professor Susan Jebb and Dr. Brian Cook at the Nuffield Department of Primary Care Health Sciences at the University of Oxford.
FUNDING/SUPPORT:
This work was supported by the National Institute of Aging of the National Institutes of Health (grant number 1K01AG064146-01).
ROLE OF THE FUNDER/SPONSOR:
The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Footnotes
Trial registration: This study is registered at www.clinicaltrials.gov NCT04766034.
CONFLICT OF INTEREST DISCLOSURES: None reported.
ACCESS TO DATA AND DATA ANALYSIS:
Dr. Rummo had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
REFERENCES
- 1.Hu EA, Steffen LM, Coresh J, Appel LJ, Rebholz CM. Adherence to the Healthy Eating Index-2015 and Other Dietary Patterns May Reduce Risk of Cardiovascular Disease, Cardiovascular Mortality, and All-Cause Mortality. J Nutr. 2020;150(2):312–321. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Zhan J, Liu YJ, Cai LB, Xu FR, Xie T, He QQ. Fruit and vegetable consumption and risk of cardiovascular disease: A meta-analysis of prospective cohort studies. Crit Rev Food Sci Nutr. 2017;57(8):1650–1663. [DOI] [PubMed] [Google Scholar]
- 3.Fang Zhang F, Liu J, Rehm CD, Wilde P, Mande JR, Mozaffarian D. Trends and Disparities in Diet Quality Among US Adults by Supplemental Nutrition Assistance Program Participation Status. JAMA Netw Open. 2018;1(2):e180237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Lee SH, Moore LV, Park S, Harris DM, Blanck HM. Adults Meeting Fruit and Vegetable Intake Recommendations - United States, 2019. MMWR Morb Mortal Wkly Rep. 2022;71(1):1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Andreyeva T, Marple K, Moore TE, Powell LM. Evaluation of Economic and Health Outcomes Associated With Food Taxes and Subsidies: A Systematic Review and Meta-analysis. JAMA Netw Open. 2022;5(6):e2214371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Moran A, Thorndike A, Franckle R, et al. Financial Incentives Increase Purchases Of Fruit And Vegetables Among Lower-Income Households With Children. Health Aff (Millwood). 2019;38(9):1557–1566. [DOI] [PubMed] [Google Scholar]
- 7.Olsho LE, Klerman JA, Wilde PE, Bartlett S. Financial incentives increase fruit and vegetable intake among Supplemental Nutrition Assistance Program participants: a randomized controlled trial of the USDA Healthy Incentives Pilot. Am J Clin Nutr. 2016;104(2):423–435. [DOI] [PubMed] [Google Scholar]
- 8.Rummo PE, Noriega D, Parret A, Harding M, Hesterman O, Elbel BE. Evaluating A USDA Program That Gives SNAP Participants Financial Incentives To Buy Fresh Produce In Supermarkets. Health Aff (Millwood). 2019;38(11):1816–1823. [DOI] [PubMed] [Google Scholar]
- 9.Verghese A, Raber M, Sharma S. Interventions targeting diet quality of Supplemental Nutrition Assistance Program (SNAP) participants: A scoping review. Prev Med. 2019;119:77–86. [DOI] [PubMed] [Google Scholar]
- 10.Lyerly R, Rummo P, Amin S, et al. Effectiveness of mobile produce markets in increasing access and affordability of fruits and vegetables among low-income seniors. Public Health Nutr. 2020;23(17):3226–3235. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Rummo PE, Lyerly R, Rose J, Malyuta Y, Cohen ED, Nunn A. The impact of financial incentives on SNAP transactions at mobile produce markets. Int J Behav Nutr Phys Act. 2021;18(1):26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Atoloye AT, Savoie-Roskos MR, Durward CM. Higher Fruit and Vegetable Intake Is Associated with Participation in the Double Up Food Bucks (DUFB) Program. Nutrients. 2021;13(8). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Engel K, Ruder EH. Fruit and Vegetable Incentive Programs for Supplemental Nutrition Assistance Program (SNAP) Participants: A Scoping Review of Program Structure. Nutrients. 2020;12(6). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Public Assistance Act. RI H7490. https://www.billtrack50.com/billdetail/1457192/31607. Accessed January 2023. [Google Scholar]
- 15.Roberto CA, Kawachi I. Behavioral economics and public health. Oxford University Press; 2015. [Google Scholar]
- 16.Perez CL, Moran A, Headrick G, McCarthy J, Cradock AL, Pollack Porter KM. State and Local Healthy Kids’ Meal Laws in the United States: A Review and Content Analysis. J Acad Nutr Diet. 2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Anzman-Frasca S, Mueller MP, Sliwa S, et al. Changes in children’s meal orders following healthy menu modifications at a regional U.S. restaurant chain. Obesity (Silver Spring). 2015;23(5):1055–1062. [DOI] [PubMed] [Google Scholar]
- 18.Coffino JA, Han GT, Evans EW, Luba R, Hormes JM. A Default Option to Improve Nutrition for Adults With Low Income Using a Prefilled Online Grocery Shopping Cart. J Nutr Educ Behav. 2021;53(9):759–769. [DOI] [PubMed] [Google Scholar]
- 19.Coffino JA, Udo T, Hormes JM. Nudging while online grocery shopping: A randomized feasibility trial to enhance nutrition in individuals with food insecurity. Appetite. 2020;152:104714. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Carroll KA, Samek A, Zepeda L. Food bundling as a health nudge: Investigating consumer fruit and vegetable selection using behavioral economics. Appetite. 2018;121:237–248. [DOI] [PubMed] [Google Scholar]
- 21.Guenther PM, Dodd KW, Reedy J, Krebs-Smith SM. Most Americans eat much less than recommended amounts of fruits and vegetables. J Am Diet Assoc. 2006;106(9):1371–1379. [DOI] [PubMed] [Google Scholar]
- 22.Leung CW, Musicus AA, Willett WC, Rimm EB. Improving the Nutritional Impact of the Supplemental Nutrition Assistance Program:: Perspectives From the Participants. American Journal of Preventive Medicine. 2017;52(2):S193–S198. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Trude ACB, Lowery CM, Ali SH, Vedovato GM. An equity-oriented systematic review of online grocery shopping among low-income populations: implications for policy and research. Nutr Rev. 2022;80(5):1294–1310. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Duffy EW, Lo AE, Hall MG, Taillie LS, Ng SW. Prevalence and Demographic Correlates of Online Grocery Shopping: Results from a Nationally Representative Survey During the COVID-19 Pandemic. Public Health Nutr. 2022:1–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Jones JW. COVID-19 Working Paper: Supplemental Nutrition Assistance Program and Pandemic Electronic Benefit Transfer Redemptions during the Coronavirus Pandemic. 2021. [Google Scholar]
- 26.Berinsky AJ, Huber GA, Lenz GSJPa. Evaluating online labor markets for experimental research: Amazon. com’s Mechanical Turk. 2012;20(3):351–368. [Google Scholar]
- 27.Jeong M, Zhang D, Morgan JC, et al. Similarities and Differences in Tobacco Control Research Findings From Convenience and Probability Samples. Ann Behav Med. 2019;53(5):476–485. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Weinberg JD, Freese J, McElhattan DJSS. Comparing data characteristics and results of an online factorial survey between a population-based and a crowdsource-recruited sample. 2014;1. [Google Scholar]
- 29.Roberto CA, Kawachi I. Use of psychology and behavioral economics to promote healthy eating. American Journal of Preventive Medicine. 2014;47(6):832–837. [DOI] [PubMed] [Google Scholar]
- 30.Epstein LH, Finkelstein E, Raynor H, et al. Experimental analysis of the effect of taxes and subsides on calories purchased in an on-line supermarket. Appetite. 2015;95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Rummo PE, Roberto CA, Thorpe LE, Troxel AB, Elbel B. Age-Specific Differences in Online Grocery Shopping Behaviors and Attitudes among Adults with Low Income in the United States in 2021. Nutrients. 2022;14(20). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.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). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Rouhani MH, Haghighatdoost F, Surkan PJ, Azadbakht L. Associations between dietary energy density and obesity: A systematic review and meta-analysis of observational studies. Nutrition. 2016;32(10):1037–1047. [DOI] [PubMed] [Google Scholar]
- 34.Tiehen L, Newman C, Kirlin JA. The food-spending patterns of households participating in the Supplemental Nutrition Assistance Program: Findings from USDA’s FoodAPS. 2017. [Google Scholar]
- 35.Ogden CL, Carroll MD, Fryar CD, Flegal KM. Prevalence of Obesity Among Adults and Youth: United States, 2011–2014. NCHS Data Brief. 2015(219):1–8. [PubMed] [Google Scholar]
- 36.Hall MG, Higgins ICA, Grummon AH, 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). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Visscher TL, Seidell JC. The public health impact of obesity. Annual Review of Public Health. 2001;22:355–375. [DOI] [PubMed] [Google Scholar]
- 38.Leung CW, Ding EL, Catalano PJ, Villamor E, Rimm EB, Willett WC. Dietary intake and dietary quality of low-income adults in the Supplemental Nutrition Assistance Program. Am J Clin Nutr. 2012;96(5):977–988. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Andreyeva T, Tripp AS, Schwartz MB. Dietary Quality of Americans by Supplemental Nutrition Assistance Program Participation Status: A Systematic Review. Am J Prev Med. 2015;49(4):594–604. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Hiza HA, Casavale KO, Guenther PM, Davis CA. Diet quality of Americans differs by age, sex, race/ethnicity, income, and education level. J Acad Nutr Diet. 2013;113(2):297–306. [DOI] [PubMed] [Google Scholar]
- 41.Ducrot P, Julia C, Méjean C, et al. Impact of Different Front-of-Pack Nutrition Labels on Consumer Purchasing Intentions: A Randomized Controlled Trial. Am J Prev Med. 2016;50(5):627–636. [DOI] [PubMed] [Google Scholar]
- 42.Epstein LH, Finkelstein EA, Katz DL, Jankowiak N, Pudlewski C, Paluch RA. Effects of nutrient profiling and price changes based on NuVal® scores on food purchasing in an online experimental supermarket. Public Health Nutr. 2016;19(12):2157–2164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Shin S, Chakraborty B, Yan X, van Dam RM, Finkelstein EA. Evaluation of Combinations of Nudging, Pricing, and Labeling Strategies to Improve Diet Quality: A Virtual Grocery Store Experiment Employing a Multiphase Optimization Strategy. Ann Behav Med. 2022. [DOI] [PubMed] [Google Scholar]
- 44.Egnell M, Boutron I, Péneau S, et al. Randomised controlled trial in an experimental online supermarket testing the effects of front-of-pack nutrition labelling on food purchasing intentions in a low-income population. BMJ Open. 2021;11(2):e041196. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Koutoukidis DA, Jebb SA, Ordóñez-Mena JM, 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. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Breugelmans E, Campo K, Gijsbrechts EJML. Shelf sequence and proximity effects on online grocery choices. 2007;18(1):117–133. [Google Scholar]
- 47.Breugelmans E, Campo KJJor. Effectiveness of in-store displays in a virtual store environment. 2011;87(1):75–89. [Google Scholar]
- 48.Epstein LH, Finkelstein E, Raynor H, et al. Experimental analysis of the effect of taxes and subsides on calories purchased in an on-line supermarket. Appetite. 2015;95:245–251. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Mozaffarian D, Liu J, Sy S, et al. Cost-effectiveness of financial incentives and disincentives for improving food purchases and health through the US Supplemental Nutrition Assistance Program (SNAP): A microsimulation study. PLoS Med. 2018;15(10):e1002661. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Bartlett S, Associates A. Evaluation of the Healthy Incentives Pilot (HIP), final report. United States Department of Agriculture, Food and Nutrition Service, Office; …; 2014. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Dr. Rummo had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.