Gains Made By Walmart’s Healthier Food Initiative Mirror Preexisting Trends

Scholars increasingly recognize that food retailers are unique and critical allies in the fight against obesity. ^1,2 Since 2011 three of the largest grocers in the United States have implemented healthier food initiatives intended to improve the healthfulness of food that consumers purchase. These efforts most often consist of strategies to help consumers identify healthier options through shelf or front-of-package labels. Sometimes they include additional measures, such as strategic price cuts, product reformulation, or additional marketing initiatives. Such retailer-based strategies have major potential to improve the nutrient profile of what US households purchase and consume, not only because food stores in general provide the majority of daily energy for US children and adults, ^3,4 but also because national trends toward chain stores and consolidation ⁵ mean that a limited number of top retailers account for the majority of US food purchases. ⁶

Yet no independent work has evaluated whether a multicomponent healthier food initiative at a major national retailer has actually succeeded in improving the nutritional profile of food purchases. One important case study is Walmart, which is the country’s largest food retailer. ⁶ Over 80 percent of households shopped at Walmart in 2012, ⁷ and it is one of the largest recipients of Supplemental Nutrition Assistance Program spending. ⁸

In 2011 Walmart announced a healthier food initiative with the stated intent of helping consumers make healthier food purchases. ⁹ The initiative included a front-of-package labeling system to identify items that met certain nutrition criteria and strategic price reductions on healthier items. To improve the healthfulness of the food available, the initiative also had the following goals: to eliminate trans fat, reduce sodium by 25 percent, and reduce added sugar by 10 percent in key product categories by 2015. ¹⁰ Walmart’s website suggests that these goals were to be achieved through the reformulation of products, but reductions in key nutrients across these product categories could also be achieved through the introduction of new products or the removal of other products.

The goal of this study was to use a natural experiment, capitalizing on observations of households’ packaged food purchases before and after the implementation of Walmart’s healthier food initiative, to examine whether such initiatives can improve the nutritional profile of packaged food purchases at major chain food retailers. However, there are a number of major challenges associated with examining healthier food initiatives at chain food retailers, especially when those interventions are not planned experiments.

For example, the only study to our knowledge that has examined a chain retailer–based initiative focused on a shelf labeling program and found that the use of a star icon on shelf labels to indicate healthier products was linked to significant but small (less than 2 percent) increases in the percentages of purchases that had received a star icon as part of the intervention. ¹¹ However, this analysis simply examined the proportion of healthier products purchased before and after the implementation of the intervention. This type of simple pre- versus post-intervention analysis cannot determine whether changes are attributable to the intervention or to other concurrent secular trends that affected food purchasing.

In the case of Walmart, it is important to distinguish between the ostensible effects of the healthier food initiative and those of other concurrent trends, including the Great Recession, large increases in global food prices, and other shifts in the food retail environment such as industrywide changes in product assortment or formulation. ^12,13

In addition, the lack of an appropriate control group can sometimes pose a major challenge to evaluating food retailer–based interventions. For example, two studies, one in the United States ¹⁴ and the other in Scotland, ² examined the impact of introducing a food retailer into food deserts on fruit and vegetable intake. The studies used controlled quasi-experimental designs, in which the intervention community was matched to a control community. The studies found that the intervention was not associated with increased fruit and vegetable intake. However, a similar intervention in Britain without a comparison community did find improvements in fruit and vegetable intake. ¹⁵ Another study, by Gary Foster and coauthors, used a cluster-randomized study at the store level to examine a six-month in-store marketing intervention to promote sales of healthier products. It found that the intervention was linked with greater sales of skim and 1 percent milk, water, and two of three types of frozen meals, but not cereal, whole or 2 percent milk, or regular or diet beverages. ¹⁶

In the case of Walmart, no clear comparison group exists. To the best of our knowledge, the intervention occurred in all Walmart stores at the same or nearly the same time, and no stores remained unexposed to the intervention.

Ideally, we would observe the counterfactual: What would the nutritional profile of purchases look like if the healthier food initiative had not been implemented, holding all other factors constant? This approach was recently used by two of the authors, Shu Wen Ng and Barry Popkin, to evaluate the Healthy Weight Commitment Foundation’s pledge to cut 1.5 trillion calories from the US food supply. ¹⁷ However, it has not yet been applied to examine similar pledges by food retailers.

These evaluation efforts are further complicated by concerns about selection bias issues: As the retailer implements its healthier food initiative, it may attract more customers who are seeking out these new healthier food options. Thus, a key question is whether a retailer truly improved the nutrient profile of packaged food purchases or whether it simply attracted a more health-conscious consumer base.

The final challenge is determining the appropriate baseline period against which to measure these changes. Since in a natural experiment investigators do not implement an intervention but observe variation in response to natural events, it is not always obvious when the intervention occurred.

In the present example, Walmart formally announced its healthier food initiative in January 2011. However, much of its online marketing suggests an earlier start date (for example, “between 2008 and 2011, we decreased sodium by 13% across the commercial bread category. This is equivalent to removing more than 1.5 million pounds of salt from the market baskets of our shoppers”). ¹⁰ While it is not uncommon for organizations to initiate changes in products before the formal beginning of a program or policy, as demonstrated by recent school lunch and menu labeling initiatives and many corporate voluntary initiatives, ^18–20 this ambiguity surrounding the implementation date is often overlooked.

The objective of this study was to examine whether Walmart’s healthier food initiative improved the nutritional profile of packaged food purchases from before the implementation of the initiative (2000–10) to after the implementation (2011–13), including the density of nutrients such as energy, sodium, and total sugar (that is, naturally occurring sugar plus sugar added in processing), as well as the contributions of key food groups. We conducted counterfactual simulations by comparing the projected pre-implementation trends in the nutrient profile of Walmart packaged food purchases to the observed post-implementation trends in that profile. We examined whether those results varied using the official launch date of the initiative or a date indicated by the data as the cut-off between the pre- and post-implementation periods. We also compared trends at Walmart to concurrent trends in packaged food purchases at other chain retailers (grocery stores, supermarkets, and supercenters with at least ten locations) to examine whether the healthier food initiative was associated with any changes above and beyond industry trends.

Study Data And Methods

Data

This study used data on households’ packaged food purchases from 2000 to 2013 from the Nielsen Homescan longitudinal data set. Details about the sampling frame and methods have been published elsewhere. ²¹ In brief, households were sampled from seventy-six metropolitan and nonmetropolitan US markets, and members of the participating households used a handheld scanner to record information on all packaged food purchases. Packaged food purchases included all food and beverages with a Universal Product Code (bar code), including all consumer packaged goods and packaged fresh fruit and vegetables (for example, a bag of lemons) but excluding unpackaged meat and produce (such as a single lemon).

To understand whether shifts in the nutrient profile of Walmart packaged food purchases were above and beyond industry trends, we also examined changes in packaged food purchases from other chain retailers, which are the most comparable group of stores given their size and product assortment. Information on packaged food purchases was linked at the bar-code level to nutrition data from the Nutrition Facts panel ²² and updated annually; it included both nationally branded and store-brand products. However, contract agreements with our data vendors meant that we were unable to identify and examine store-brand products sold by specific retailers. Household-quarter observations were included for households in the sample that made packaged food purchases at Walmart ( n = 1,212,803) or other chain retailers ( n = 2,521,128) for at least two quarters during the study period. Demographic characteristics of the sample are in online Appendix Exhibit 1. ²³

Nutrition Outcomes

Nutrition outcomes were evaluated separately for packaged food purchases from Walmart and other chain retailers and included density of energy (kcal/100 g), total sugar (g/100 g), saturated fat (g/100 g), and sodium (mg/100 g). A volume-based (per 100 grams) relative measure was required to control for changes in the overall amount of packaged food purchases purchased over time. We also examined shifts in percentage volume purchased (percentage of grams purchased) from key food groups, including sugar-sweetened beverages, grain-based desserts, savory snacks (that is, salty snack foods such as chips or pretzels), candy, fresh or frozen fruit, and fresh or frozen vegetables.

Addressing Selection Bias

To deal with the potential selection bias of shopping at a given retailer over time, we created two sets of quarterly inverse probability weights to account for the likelihood of being a Walmart shopper or other chain retailer shopper, conditional upon variables associated with shopping these retailer types. ²⁴

In addition, we used fixed-effects models to examine average within-household effects over time. In this case, fixed-effects models can correct for selection bias if the likelihood of shopping at a certain retailer is associated with fixed characteristics of the household (for example, income and race/ethnicity). In the models, each household essentially served as its own control.

Counterfactual Simulation

We first identified potential dates when the healthier food initiative occurred by using switching regression ²⁵ for both energy and sodium densities of Walmart packaged food purchases, since food industry initiatives to reduce sodium in processed food began before Walmart’s healthier food initiative. ²⁶ For energy density, switching regression revealed a shift at 2011 as well as a shift at 2007 and 2013 (Appendix Exhibit 2). ²³ Sodium density showed switches in different directions at two- and three-year intervals across the time period. Thus, for all models, we examined both 2011 and 2007 as potential healthier food initiative initiation dates, since 2011 was Walmart’s stated initiation date and 2007 was the second most plausible date, given the 2008 date indicated in some of Walmart’s marketing materials.

Using these start dates, we tested the shape of the trend. In the energy models, we found that when we used the 2011 start date, a quadratic trend provided the best fit (based on model R ² and visual inspection). In contrast, when we used the 2007 start date, a linear trend provided the best fit. For the sodium model, a linear trend provided the best fit for the 2011 start date, while a quadratic term provided the best fit for the 2007 start date. Models with the appropriate terms were used for the remainder of the analyses.

Statistical Analysis

We used fixed-effects models to separately model all nutrient outcomes at Walmart and other chain retailers. All models included covariates to control for secular trends in the economic, food retail, and price environments, including average quarterly market-level unemployment rates, ²⁷ the weighted average of prices at each retailer type, and average market-level Walmart density (Walmart supercenters per 100,000 people). ^28,29 Household-level covariates were household race/ethnicity (non-Hispanic white, Hispanic, non-Hispanic black, or non-Hispanic other), eligibility for Supplemental Nutrition Assistance Program assistance (income up to 130 percent of the federal poverty level), head of household education (no more than a high school diploma, some college, or more than a college degree), household type (single adult, multiple adults with no children, or one or more adults with one or more children) and household composition (numbers of adults ages 19–49 and 50 years or older, and numbers of children ages 0–5 and 6–18 years).

To examine the observed nutrient profile of purchases before and after the implementation of the healthier food initiative, we used Stata’s margins command to predict the mean nutrient density or percentage volume from food groups for each year for each retailer. For models examining the counterfactual post-implementation nutrient profile, we first estimated the pre-implementation time trend (2000–10 or 2000–07; linear or quadratic) for each nutrient outcome. We then projected this time trend into the post-implementation period to obtain the predicted mean nutrient outcomes that would have been observed in that period had the initiative not been implemented. We used Stata’s predict command to predict mean nutrient outcomes during the period after the initiative’s implementation.

We considered the observed mean nutrient profile significantly different from the counterfactual profile if 99 percent confidence intervals did not overlap ( $p<0.01$ ). All statistical analyses were performed using Stata, version 13.

Limitations

Several limitations meant that we were not able to detect all potential changes related to the healthier food initiative. For example, we could not examine the impact of specific components of the initiative, nor could we identify the key categories specifically designated by Walmart for reformulation, as this is not public information. We also could not identify certain nutrients, such as trans fat, which did not have to appear on nutrition labels before 2006; added sugar, which under current regulations is not required to appear on the nutrition facts panel; or other nutrients such as fiber, calcium, or iron.

In addition, Nielsen Homescan does not capture products without a bar code, including loose produce. Although unpackaged produce accounted for only approximately 2–3 percent of Walmart food expenditures in our sample, we likely underestimated the effects of Walmart’s efforts to increase the availability and affordability of produce ³⁰ on fruit and vegetable purchases.

Finally, we could not distinguish in our analyses between Walmart’s store-brand products—which were a major focus of the company’s efforts in product reformulation and labeling—and other products purchased at Walmart. Had we been able to analyze Walmart’s store-brand products separately, we might have observed larger changes in the nutritional profile of Walmart packaged food purchases.

Study Results

Walmart packaged food purchases showed a 74 kcal/100g decline in energy density from 2000 to 2013, although the decline leveled off around the late 2000s ( Exhibit 1 ). In fact, after the implementation of the healthier food initiative beginning in 2011, Walmart packaged food purchases declined less in energy density than would be expected based on the trend before 2011.

SOURCE Calculations based in part on data reported by Nielsen through its Homescan Services for the food and beverage categories for the US market. Copyright © 2015, The Nielsen Company. ^a The observed nutritional profile of post-HFI purchases differs from the 2007 counterfactual based on the pre-HFI trend ( $p<0.01$ ). ^b The observed nutritional profile of post-HFI purchases differs from the 2011 counterfactual based on the pre-HFI trend ( $p<0.01$ ).

The rate of decline for total sugar density of Walmart packaged food purchases also slowed around 2007 ( Exhibit 2 ). The total sugar density of packaged food purchases during the period after 2011 was very similar to what we would have expected based on trends before 2011, and in 2013 the total sugar density was slightly less than expected based on those trends.

SOURCE Calculations based in part on data reported by Nielsen through its Homescan Services for the food and beverage categories for the US market. Copyright © 2015, The Nielsen Company. ^a The observed nutritional profile of post-HFI purchases differs from the 2007 counterfactual based on the pre-HFI trend ( $p<0.01$ ). ^b The observed nutritional profile of post-HFI purchases differs from the 2011 counterfactual based on the pre-HFI trend ( $p<0.01$ ).

The saturated fat density of Walmart packaged food purchases showed small declines from 2000 to 2013 ( Exhibit 3 ). Nonetheless, the saturated fat density after 2011 was slightly but significantly higher than expected based on the trend before 2011.

SOURCE Calculations based in part on data reported by Nielsen through its Homescan Services for the food and beverage categories for the US market. Copyright © 2015, The Nielsen Company. ^a The observed nutritional profile of post-HFI purchases differs from the 2011 counterfactual based on the pre-HFI trend ( $p<0.01$ ).

The sodium density of Walmart packaged food purchases decreased by 33 mg/100g from 2000 to 2013 ( Exhibit 4 ). However, the sodium density after 2011 was virtually identical to what we would have expected based on trends before 2011.

SOURCE Calculations based in part on data reported by Nielsen through its Homescan Services for the food and beverage categories for the US market. Copyright © 2015, The Nielsen Company.

Using 2007 as the year when the healthier food initiative began tended to show that post-implementation declines were even smaller than expected compared to when 2011 was used as the implementation date. In other words, the observed nutrient densities for Walmart packaged food purchases after 2007 tended to be even higher than we would have expected based on the trend before 2007 ( Exhibits 1 – 4 ). For example, the disparity between the observed decline in energy density of Walmart packaged food purchases and what we would have expected based on trends before 2007 widened ( Exhibit 1 ). And using 2007 as the start date for the healthier food initiative reversed the result for total sugar density, so that total sugar density in 2013 was substantially higher, instead of lower, than what would be expected based on trends before the initiative ( Exhibit 2 ).

Appendix Exhibit 3 ²² shows the percentage volumes purchased from key food groups at Walmart. Overall, the percentage volume of sugar-sweetened beverages purchased from Walmart declined only slightly, although this decline was more than expected based on the trend before 2011. Percentage volumes of fruit and vegetables increased only slightly, and increases during the period after implementation were the same as would have been expected based on earlier trends. Percentage volumes of grain-based desserts, savory snacks, and candy purchased from Walmart also declined, although the declines after implementation of the initiative were also the same as would have been expected based on earlier trends.

Our results were similar when we examined shifts in percentage volumes of food groups using the 2007 start date, with a reversal or widening of the gap. In other words, percentage volumes changed less than would have been expected. For example, when we used 2007 as the initiative’s start date, we found no difference between the percentage volume of sugar-sweetened beverages purchased from Walmart after the implementation of the initiative and what would have been expected based on earlier trends.

When we looked at packaged food purchases from other chain retailers, we found similar but smaller trends in nutrient densities and percentage volumes from key food groups (Appendix Exhibits 4 and 5). ²³ For energy, total sugar, and sodium densities, declines after the implementation of Walmart’s healthier food initiative were larger than would have been expected based on trends before the implementation ( $p<0.01$ ). However, the overall declines were minimal (11 kcal/100g for energy, <2 g/100g for total sugar, and 20 mg/100g for sodium).

Discussion

From 2000 to 2013, energy, total sugar, and sodium densities of packaged food purchases from Walmart declined substantially, and these changes were greater than those observed in packaged food purchases from other chain food retailers. On the whole, however, using 2011 as the date for the implementation of Walmart’s healthier food initiative, we found that shifts in nutrient density and percentage volumes of key food groups after the implementation were similar to or less than what we would have expected had previous trends simply continued.

These results are contrary to what we would expect to find if the healthier food initiative truly marked a turning point in how Walmart formulated, priced, and marketed its foods. Had that been the case, we would expect to find that nutrient densities of Walmart packaged food purchases declined more quickly after the initiative was implemented.

Instead, we found that the declines in nutrient density and shifts in percentage volumes of key food groups were steepest in the early 2000s and leveled off around 2007 or 2008. In fact, these results mirrored those of the Healthy Weight Commitment Foundation evaluation, which found that declines in caloric purchases after food companies’ pledged to remove a trillion calories from the food supply were smaller than expected based on the pre-pledge trend. ¹⁷ In our study, declines in the earlier 2000s, followed by a leveling off around 2007, could simply reflect Walmart’s continued expansion of its grocery lines within its supercenters and increased consumer perception of Walmart as a place to shop for groceries, not just the occasional snack or soda. By the end of the observation period, purchases at Walmart simply resembled those at other chain retailers, which suggests that Walmart’s improvements brought it up to par with other retailers instead of representing an improvement beyond current industry standards.

One notable exception was the shift in percentage volume of sugar-sweetened beverage purchases from Walmart. Our results show that before implementation of the healthier food initiative, the percentage volume of these beverages was increasing. After implementation, purchases of the beverages declined by 1 percent. Although this decline was greater than expected based on pre-implementation trends, it is not clear whether this decline was driven by the healthier food initiative, as Walmart has not publicly indicated whether its initiative targeted sugar-sweetened beverage purchases. It is possible the observed decline in beverage purchases simply reflected a general downward trend in the purchase of sugar-sweetened beverages in the United States. Recent evidence has shown that purchases of the beverages have declined in recent years, ³¹ possibly because of shifting public attitudes toward sugar-sweetened beverages as a result of public health campaigns about the beverages’ potential health consequences. However, such a shift in awareness would have likely resulted in a downward shift in sugar-sweetened beverage purchases at other chain retailers as well, which we did not see.

It is possible that we would have found changes after implementation if we had been able to isolate the effects of Walmart’s healthier food initiative on its store-brand products—which were a focus of the company’s efforts. However, we were unable to analyze Walmart’s store-brand products separately. If changes in store-brand products had been large enough, they would have been reflected in our analysis of overall purchases.

This study also shows that the choice of baseline period matters for determining whether healthier food initiatives are responsible for observed changes in nutritional profile. When we used a 2007 start date for Walmart’s initiative, we found that declines in the nutrient densities of Walmart packaged food purchases were markedly less than we would have expected. This likely occurred because for the energy-containing nutrients, the majority of declines occurred before 2007 and then leveled off. Thus, when we used 2011 as the start date, the quadratic trend fit best and created a trajectory after the initiative that was very similar to what we actually observed for energy, total sugar, and saturated fat densities.

However, a 2007 start date gave a linear trajectory after the initiative. These declines were occurring at a relatively steeper rate before 2007. Thus, when we used that year as the start date and looked at what happened after it, we found that the nutrient densities declined significantly less after the implementation of the initiative than we would have expected had the trend before the implementation continued. Future evaluations, especially those relying on natural experiments in which the implementation date is not always clear, should consider how results change when different baseline periods are used.

It was also surprising that although the energy, total sugar, and sodium density of packaged food purchases from other chain retailers declined more rapidly than expected during the period after the implementation of Walmart’s healthier food initiative, overall declines were minimal. Even though the time trend was based on the implementation of Walmart’s initiative, we would have expected to see larger declines, especially after 2007 and 2011, in part because of Walmart’s influence. Because of Walmart’s size, its price cuts or negotiations with suppliers to reformulate products could have led to lower prices for produce or the introduction of healthier options at other chain retailers as well. ^32–34 Additionally, we might have expected to see steeper declines in later years as a result of the implementation of healthier food initiatives at other major retailers such as Safeway, Food Lion, and Kroger, ^35–37 although the geographic scope of these initiatives is currently unclear.

Conclusion

From 2000 to 2013, Walmart, the largest food retailer in the United States, showed major declines in the energy, total sugar, and sodium density of packaged food purchases as well as in percentage volumes purchased from sugary beverages, grain-based desserts, snacks, and candy. These declines were above and beyond concurrent trends in purchases at comparable food retailers. However, the majority of these changes occurred before the official implementation of Walmart’s healthier food initiative, which indicates that the initiative was not responsible for these improvements. The limitations of this study, including our inability to analyze Walmart store-brand products separately from other products purchased at Walmart, suggest that further research into Walmart’s initiative as well as other food retailer–based initiatives is warranted.

Nonetheless, our results suggest that retailer-based initiatives may not result in meaningful changes in the nutritional profile of what people buy. In fact, these results are comparable to those from recent work published in the food access literature, which shows that placing food retailers in an areas of low access is not enough to improve the healthfulness of food purchases. ^14,38,39 It seems as though simply changing the food retail environment is not enough. In fact, a recent study concluded that educational differences, not the introduction of new stores or changes in store offerings, were the major driver of disparities in the nutritional quality of food purchases. ³⁹ Thus, while food retailers should be engaged in efforts to create a healthier food environment, more systemic shifts in the underlying characteristics and preferences of the population may be needed to meaningfully improve the healthfulness of food purchases.

ACKNOWLEDGMENTS

Funding for this study came from the Robert Wood Johnson Foundation (Grant Nos. 67506, 68793, 70017, and 71837), the National Institutes of Health (Grant No. R01DK098072), and the Carolina Population Center (Grant Nos. R24 HD050924 and T32 HD007168). The authors thank Donna Miles for exceptional assistance with the data management and programming; Tom Swasey for graphics support; and Mary Story, David Guilkey, and Anna Maria Siega-Riz for their review of our analysis and insightful feedback.

NOTES

• 1 Glanz K , Bader MD , Iyer S . Retail grocery store marketing strategies and obesity: an integrative review . Am J Prev Med . 2012 ; 42 ( 5 ): 503 – 12 . Crossref, Medline, Google Scholar
• 2 Cummins S , Petticrew M , Higgins C , Findlay A , Sparks L . Large scale food retailing as an intervention for diet and health: quasi-experimental evaluation of a natural experiment . J Epidemiol Community Health . 2005 ; 59 ( 12 ): 1035 – 40 . Crossref, Medline, Google Scholar
• 3 Poti JM , Popkin BM . Trends in energy intake among US children by eating location and food source, 1977–2006 . J Am Diet Assoc . 2011 ; 111 ( 8 ): 1156 – 64 . Crossref, Medline, Google Scholar
• 4 Smith LP , Ng SW , Popkin BM . Trends in US home food preparation and consumption: analysis of national nutrition surveys and time use studies from 1965–1966 to 2007–2008 . Nutr J . 2013 ; 12 ( 1 ): 45 . Crossref, Medline, Google Scholar
• 5 Wood S . Revisiting the US food retail consolidation wave: regulation, market power, and spatial outcomes . J Econ Geogr . 2013 ; 13 ( 2 ): 299 – 326 . Crossref, Google Scholar
• 6 Department of Agriculture, Economic Research Service . Retailing and wholesaling: retail trends [Internet]. Washington (DC) : USDA ; [last updated 2015 May 11; cited 2015 Aug 28 ]. Available from: http://www.ers.usda.gov/topics/food-markets-prices/retailing-wholesaling/retail-trends.aspx Google Scholar
• 7 Taillie LS , Ng SW , Popkin BM . “Big box” stores abound: a review of the global growth of supercenters and the potential impact on human health and nutrition . Nutr Rev . Forthcoming. Google Scholar
• 8 Banjo S , Gasparro A . Retailers brace for reduction in food stamps . Wall Street Journal . 2014 Nov 4 . Google Scholar
• 9 Walmart [Internet]. Bentonville (AR) : Walmart . News release, Walmart launches major initiative to make food healthier and healthier food more affordable ; 2011 Jan 20 [cited 2015 Aug 28 ]. Available from: http://news.walmart.com/news-archive/2011/01/20/walmart-launches-major-initiative-to-make-food-healthier-healthier-food-more-affordable Google Scholar
• 10 Wal-Mart . Making healthier food a reality for all [Internet]. Bentonville (AR) : Walmart ; c 2015 [cited 2015 Aug 28 ]. Available from: http://corporate.walmart.com/global-responsibility/hunger-nutrition/our-commitments Google Scholar
• 11 Sutherland LA , Kaley LA , Fischer L . Guiding stars: the effect of a nutrition navigation program on consumer purchases at the supermarket . Am J Clin Nutr . 2010 ; 91 ( 4 ): 1090S – 4S . Crossref, Medline, Google Scholar
• 12 Brinkman HJ , de Pee S , Sanogo I , Subran L , Bloem MW . High food prices and the global financial crisis have reduced access to nutritious food and worsened nutritional status and health . J Nutr . 2010 ; 140 ( 1 ): 153S – 61S . Crossref, Medline, Google Scholar
• 13 Beatty TKM , Senauer B . The new normal? U.S. food expenditure patterns and the changing structure of food retailing . Am J Agric Econ . 2012 ; 95 ( 2 ): 318 – 24 . Crossref, Google Scholar
• 14 Cummins S , Flint E , Matthews SA . New neighborhood grocery store increased awareness of food access but did not alter dietary habits or obesity . Health Aff (Millwood) . 2014 ; 33 ( 2 ): 283 – 91 . Go to the article, Google Scholar
• 15 Wrigley N , Warm D , Margetts B . Deprivation, diet, and food-retail access: findings from the Leeds “food deserts” study . Environment and Planning . 2003 ; 35 ( 1 ): 151 – 88 . Crossref, Google Scholar
• 16 Foster GD , Karpyn A , Wojtanowski AC , Davis E , Weiss S , Brensinger C , et al. Placement and promotion strategies to increase sales of healthier products in supermarkets in low-income, ethnically diverse neighborhoods: a randomized controlled trial . Am J Clin Nutr . 2014 ; 99 ( 6 ): 1359 – 68 . Crossref, Medline, Google Scholar
• 17 Ng SW , Popkin BM . The Healthy Weight Commitment Foundation pledge: calories purchased by U.S. households with children, 2000–2012 . Am J Prev Med . 2014 ; 47 ( 4 ): 520 – 30 . Crossref, Medline, Google Scholar
• 18 Ohri-Vachaspati P , Turner L , Chaloupka FJ . Fresh Fruit and Vegetable Program participation in elementary schools in the United States and availability of fruits and vegetables in school lunch meals . J Acad Nutr Diet . 2012 ; 112 ( 6 ): 921 – 6 . Crossref, Medline, Google Scholar
• 19 Bassett MT , Dumanovsky T , Huang C , Silver LD , Young C , Nonas C , et al. Purchasing behavior and calorie information at fast-food chains in New York City, 2007 . Am J Public Health . 2008 ; 98 ( 8 ): 1457 – 9 . Crossref, Medline, Google Scholar
• 20 Ng SW , Slining MM , Popkin BM . The Healthy Weight Commitment Foundation pledge: calories sold from U.S. consumer packaged goods, 2007–2012 . Am J Prev Med . 2014 ; 47 ( 4 ): 508 – 19 . Crossref, Medline, Google Scholar
• 21 Zhen C , Taylor JL , Muth MK , Leibtag E . Understanding differences in self-reported expenditures between household scanner data and diary survey data: a comparison of Homescan and consumer expenditure survey . Appl Econ Perspect Policy . 2009 ; 31 ( 3 ): 470 – 92 . Google Scholar
• 22 Slining MM , Ng SW , Popkin BM . Food companies’ calorie-reduction pledges to improve U.S. diet . Am J Prev Med . 2013 ; 44 ( 2 ): 174 – 84 . Crossref, Medline, Google Scholar
• 23 To access the Appendix, click on the Appendix link in the box to the right of the article online.
• 24 Hogan JW , Lancaster T . Instrumental variables and inverse probability weighting for causal inference from longitudinal observational studies . Stat Methods Med Res . 2004 ; 13 ( 1 ): 17 – 48 . Crossref, Medline, Google Scholar
• 25 Akin JS , Guilkey DK , Popkin BM . The school lunch program and nutrient intake: a switching regression analysis . Am J Agric Econ . 1983 ; 65 ( 3 ): 477 – 85 . Crossref, Google Scholar
• 26 Dötsch M , Busch J , Batenburg M , Liem G , Tareilus E , Mueller R , et al. Strategies to reduce sodium consumption: a food industry perspective . Crit Rev Food Sci Nutr . 2009 ; 49 ( 10 ): 841 – 51 . Crossref, Medline, Google Scholar
• 27 Bureau of Labor Statistics . Local area unemployment statistics [Internet]. Washington (DC) : Department of Labor ; [cited 2015 Aug 31 ]. Available from: http://www.bls.gov/lau/ Google Scholar
• 28 Holmes TJ . Opening dates of Wal-Mart stores and supercenters, 1962–Jan 31, 2006 (Alaska and Hawaii excluded) . Minneapolis (MN) : University of Minnesota ; 2010 Mar 3 [cited 2015 Aug 31 ]. Available from: http://www.econ.umn.edu/~holmes/data/WalMart/store_openings.html Google Scholar
• 29 AggData . Complete list of Walmart locations [Internet]. Dupont (WA) : AggData ; [cited 2015 Aug 31 ]. Available from: http://www.aggdata.com/aggdata/complete-list-walmart-locations Google Scholar
• 30 Walmart . Sustainable food [Internet]. Bentonville (AR) : Walmart ; c 2015 [cited 2015 Sep 18 ]. Available from: http://corporate.walmart.com/global-responsibility/environment-sustainability/sustainable-agriculture Google Scholar
• 31 Kit BK , Fakhouri TH , Park S , Nielsen SJ , Ogden CL . Trends in sugar-sweetened beverage consumption among youth and adults in the United States: 1999–2010 . Am J Clin Nutr . 2013 ; 98 ( 1 ): 180 – 8 . Crossref, Medline, Google Scholar
• 32 Volpe RJ , Lavoie N . The effect of Wal-Mart supercenters on grocery prices in New England . Appl Econ Perspect Policy . 2008 ; 30 ( 1 ): 4 – 26 . Google Scholar
• 33 Leibtag E , Barker C , Dutko P . How much lower are prices at discount stores? An examination of retail food prices [Internet]. Washington (DC) : US Department of Agriculture, Economic Research Service ; 2010 Oct [cited 2015 Aug 31 ]. (Economic Research Report No. 105). Available from: http://www.ers.usda.gov/media/135223/err105.pdf Google Scholar
• 34 Basker E , Noel M . The evolving food chain: competitive effects of Wal-Mart’s entry into the supermarket industry . J Econ Manage Strat . 2009 ; 18 ( 4 ): 977 – 1009 . Crossref, Google Scholar
• 35 Safeway announces “SimpleNutrition,” an in-store shelf tag system, to help shoppers find the right nutrition choices for them . BusinessWire [serial on the Internet]. 2011 Feb 16 [cited 2015 Aug 31 ]. Available from: http://www.businesswire.com/news/home/20110215007790/en/Safeway-Announces-%E2%80%98SimpleNutrition%E2%80%99-In-Store-Shelf-Tag-System#.VD57YvldWSo Google Scholar
• 36 Guiding Stars . Our partners [Internet]. Scarborough (ME) : Guiding Stars ; c 2015 [cited 2015 Oct 9 ]. Available from: http://guidingstars.com/partners/our-partners/ Google Scholar
• 37 Kroger . Health matters: your health matters to us [Internet]. Cincinnati (OH) : Kroger ; c 2015 [cited 2015 Aug 31 ]. Available from: https://www.kroger.com/topic/health-matters-3 Google Scholar
• 38 Elbel B , Moran A , Dixon LB , Kiszko K , Cantor J , Abrams C , et al. Assessment of a government-subsidized supermarket in a high-need area on household food availability and children’s dietary intakes . Public Health Nutr . 2015 Feb 26 : 1 – 10 . [Epub ahead of print]. Google Scholar
• 39 Handbury J , Rahkovsky I , Schnell M . What drives nutritional disparities? Retail access and food purchases across the socioeconomic spectrum . Cambridge (MA) : National Bureau of Economic Research ; 2015 Apr . (NBER Working Paper No. 21126). Crossref, Google Scholar