Food shopping, home food availability, and food insecurity among customers in small food stores: an exploratory study

Persons experiencing food insecurity, compared to those who were food secure, shopped more frequently at small stores such as convenience and corner stores.

Abstract

Customers who frequently shop in small food stores (e.g., convenience stores) may face numerous challenges to procuring healthful food for their household, and these may vary by food security status. The purpose of this study is to examine associations between food security and food shopping-related behaviors among frequent shoppers at small stores. Our sample included participants from customer intercept interviews at small food stores in an urban area. A follow-up in-home visit with a subset of customers who reported frequently shopping in these stores (≥1/week; n = 78) included a survey and researcher-administered home food inventory. Food security status was identified via the U.S. Household Food Security Survey Six-Item Short Form. Outcomes included shopping frequency and money spent by store type (e.g., small vs. large), home-to-store distance, and observed home availability of fruits, vegetables, and obesogenic foods. We estimated associations between food security status and each outcome, adjusting for demographic and poverty-related confounders. Participants were 56% female and 65% people of color; 45% received Supplemental Nutrition Assistance Program benefits and 54% experienced food insecurity in the past year. Unadjusted models indicated several significant associations: compared to those who were food secure, food-insecure participants shopped for food/beverages at small stores more times per month, spent more on food/beverages at dollar stores, and had less home availability of fruit and obesogenic foods. Associations remained significant (p = .04) between food insecurity and shopping frequency in adjusted models. Interventions requiring or incentivizing small food stores to stock healthful products could be important for improving access to nutritious food for food-insecure persons.

Implications.

Practice: Small food stores (e.g., convenience stores and corner stores) are important venues for interventions that can help persons with food insecurity secure healthy, affordable food.

Policy: Communities should consider implementing policies that require, assist, and/or incentivize small food stores to stock healthy food and beverage options at affordable prices.

Research: Future research is needed to examine specific interventions at small food stores to both reduce food insecurity and improve healthy food offerings.

Introduction

In 2017, 12.7% of U.S. households reported experiencing food insecurity. This suggests that approximately 40 million people live in households that are uncertain of having or unable to acquire enough food to meet the needs of all household members due to insufficient money or resources for food [1]. Food insecurity is associated with a variety of negative outcomes, including poor diet quality [2–4], poor physical and mental health [5–8], and increases in health care expenditures [9].

There is increasing interest in examining how changing the community food environment can help persons experiencing food insecurity access healthy food. Exploratory research to date has aimed to understand whether food shopping behaviors and accessibility to food differ by food security status. Several studies suggest that food-insecure persons rely on small food stores, such as corner and convenience stores, for at least some of their grocery needs [10–12], and Bonanno and Li [13] concluded that improving access to small grocery/convenience stores in metropolitan areas can help mitigate the likelihood of adults experiencing food insecurity. However, reliance on small food stores can present challenges for customers given that these stores are unreliable sources of healthful foods and usually stock an abundance of unhealthy food and beverages [14, 15]. Persons who frequently shop at small stores may also differ in notable ways from other types of food shoppers, and little is known about whether similar relationships between food security status and shopping behaviors will be observed in this specific subset of the population.

Higher utilization of small food stores by food-insecure persons could partly be a consequence of living in underserved areas where spatial access to supermarkets and grocery stores that sell healthy food is limited. Although research shows that low-income populations and/or those receiving governmental food assistance (i.e., Supplemental Nutrition Assistance Program [SNAP]) often do not live in close proximity to a supermarket [16], studies examining whether this is true when comparing food-secure and food-insecure populations are inconclusive. Analyses to date have examined both home-to-store distance of the nearest supermarket and to the primary store used by a household for food acquisition. Studies of the relationships between food security status and geographic distance from home to the nearest supermarket have been mixed [17–23]. Results of how distance from home to the primary store used for household food shopping varies by food security status have also been mixed [11, 20–23]. It is unclear whether food security status may be associated with home-to-store distances for small food stores that are perhaps not the primary store used for household food shopping but still may be visited frequently for food and beverage purchases [17].

Households experiencing food insecurity are not only at risk for having less availability of food in their homes but may be particularly at risk for having less healthful options. There is some evidence that food-insecure households are less likely to have foods such as fresh fruit and vegetables, but results are inconsistent [24–27]. Most studies have also assessed home food availability by self-report rather than direct observation and/or used nonvalidated instruments [24–26].

It is difficult to compare results and make conclusive statements about the literature cited above because not only is it relatively sparse and the findings mixed but also the studies utilized varying designs, sample populations, and analytic methods. In particular, only some of the studies considered economic confounders, such as household income and governmental food assistance (i.e., SNAP), in analyses. How these economic factors are associated with food security status is important to consider and is also complex. Being food insecure is likely a product of having a lower income, and households who apply for government assistance are more likely to be food insecure. Furthermore, food insecurity could potentially be offset once a household receives SNAP benefits [28].

In this exploratory study, we aimed to address some of the gaps in the literature by assessing how food insecurity was associated with shopping frequency and money spent at different types of stores and home-to-store distances, as well as healthy and unhealthy home food availability. We also consider the important confounding role household income and SNAP participation may play in these associations. Our sample was drawn from frequent small food store shoppers in a large urban area. Results can help inform policies and interventions aimed to address food insecurity and healthy food access, particularly among those who regularly shop at small food stores.

METHODS

Data were collected as part of a large research study assessing the Staple Foods Ordinance in Minneapolis, MN, that requires minimal stocking requirements of fresh and staple foods in all food stores. Small nontraditional food stores (i.e., corner stores, convenience stores, gas stations, dollar stores, and pharmacies) are the main target of the Minneapolis policy because larger traditional stores likely already meet minimum stocking requirements and the city council recognizes that small stores can serve as sources for healthy foods for a potentially vulnerable population. The assessment of the ordinance included customer intercept interviews at small and nontraditional food stores in Minneapolis, as well as St. Paul, MN. For the current study, we analyzed data from follow-up in-home visits with a subsample of the customers from both cities who participated in intercept interviews. The Institutional Review Board of the University of Minnesota approved this study.

Study sample

Customer intercept interviews were conducted at randomly selected stores using a standard protocol (see [29] for a full description of initial store recruitment and customer intercept interview methods). In brief, we randomly selected 180 stores from 255 eligible licensed grocery retailers. We excluded supermarkets and authorized retailers in the Women, Infants, and Children (WIC) program because compliance with WIC standards would result in these stores already meeting the new ordinance standards. As customers exited stores, we collected customer intercept data at four different time points at a total of 147 stores (33 stores either refused to participate, were deemed ineligible, or went out of business). For customers to be eligible for intercept interviews, they had to speak and understand English, be at least 18 years of age, have purchased a food and/or beverage, and provide verbal informed consent. As part of the interview, we asked participants about their frequency of shopping at the store. During the first two data collection periods (n = 1,435; 668 at Time 1; 767 at Time 2), if a participant indicated shopping at least once per week at the store where they participated in the intercept interview (n = 1,055; 486 at Time 1; 569 at Time 2), they were invited to participate in future follow-up in-home visits. For the current study, we utilized data from the 78 persons who participated in the in-home visit during Time 2 (46 were recruited at Time 1 and 32 were recruited at Time 2).

In-home visits

In-home visits were conducted in 2015 by a team of two trained study staff members. The visit included a survey of demographic information and food/beverage shopping behaviors, as well as a staff-administered inventory of the observed home food environment using a validated instrument (Home Food Inventory [HFI] [30]). The HFI instrument includes approximately 200 items across 13 categories in a checklist form with yes/no (1/0) responses. Higher scores represent greater availability. All food storage areas were assessed, including the refrigerator, freezer, pantries, and cupboards.

Measures

Our exposure of interest was food security status. This was measured on the survey using the USDA Household Food Security Survey Module Six-Item Short Form (https://www.ers.usda.gov› topics › food-security-in-the-us › survey-tools). All items from the six-item module were used (food didn’t last; could not afford balanced meals; cut size of or skipped meals; ever eat less; and ever hungry), except we used the module option where responses for “cut size of or skipped meals” were “Yes, almost every month”; “Yes, some months but not every month”; “Yes, only 1 or 2 months”; or “No” (with the first two responses each counting as two affirmative responses). We dichotomized the summed score of affirmative responses (0–6) to “food secure” (score 0–1) and “food insecure” (score 2–6).

We examined 10 outcomes of food/beverage shopping behaviors from the self-administered survey. Four measures pertained to the frequency of shopping (times per month) for food and/or beverages at each of four types of stores in the past month (convenience/small grocery/gas stations, dollar stores, pharmacies, and large food stores [e.g., supermarkets, Costco]). Two measures were derived from survey items about the name and location of food stores: “the store where you buy most of food for household” and “a small store where purchased food or drinks in past week.” For these two measures, we calculated the distance (in kilometers) from the participant’s home to stores indicated using Geographic Information Systems. The last four measures were the amount spent (in dollars) in the past month on food/beverages at each of the four types of stores (convenience/small grocery/gas stations, dollar stores, pharmacies, and large food stores).

From the HFI, we utilized three validated scales of home food availability: fruits, vegetables, and obesogenic foods [30]. The fruit scale includes all fruit (canned, fresh, and frozen) and ranges from 0 to 26. The vegetable scale (range 0 to 20) includes all vegetables (canned, fresh, and frozen), including potatoes. The obesogenicity scale (range 1–71) includes regular-fat versions of cheese, milk, yogurt, other dairy, frozen desserts, prepared desserts, and savory snacks, added fats; regular-sugar beverages; processed meat; high-fat quick, microwavable foods; candy; and access to unhealthy foods in refrigerator and kitchen.

We measured several economic and demographic covariates/confounders from the survey including age (continuous), sex (male vs. female), race/ethnicity (six categories dichotomized to white vs. people of color for analysis), education (seven categories dichotomized for analysis to “some college or higher” vs. “high school or less”), employment status (full time or part time vs. unemployed/retired), household income (four categories dichotomized for analysis to “at least $25,000” vs. “under $25,000”), and number of persons in the household (dichotomized to 4+ vs. 1–3). Another measure from the survey designated whether participants received SNAP benefits in the past year (yes, no).

Analyses

We first calculated descriptive statistics for all demographic characteristics and estimated bivariate associations between each demographic characteristic and food security status. We then estimated unadjusted and adjusted regression models of associations between food security status (independent variable) and each of the food shopping and HFI outcomes (outcome variables). We estimated a series of models for each outcome variable: an unadjusted model and three adjusted models. The first adjusted model included age, sex, race, education, employment, and household size as covariates. We did not include income and SNAP status in this initial model because the relationships between these variables and food insecurity are complex. Food insecurity and SNAP assistance are both associated with being low income (see Table 1 and Appendix) and food insecurity could be mitigated by SNAP assistance. To better understand the relationships between food security status and our outcomes, given these complexities, we added household income and SNAP status in a step-wise fashion (first, income and, then, income with SNAP status). All analyses were conducted in SAS/STAT 9.4 (Cary, NC).

Table 1.

Sample characteristics

	All participants N = 78	Food secure n = 36	Food insecure n = 42
	Mean (SD)	Mean (SD)		p-valuea
Age	45 (13)	48 (15)	43 (12)	.06
	Percentage	Percentage		p-valueb
Sex: female	56	47	64	.13
Race/ethnicity				.16
Hispanic	9	14	5
Non-Hispanic
Black/African American alone	46	33	57
White alone	35	42	29
Mixed race or other	10	11	10
Education (highest level)				.45
High school degree or less	53	50	55
Some college	19	14	24
Associate degree	12	14	10
Bachelor or graduate degree	17	22	12
Employment				.04
Full time	40	46	36
Part time	23	31	17
Unemployed	30	14	43
Retired	6	9	5
Household income				.002
< 25K	51	31	69
25K−50K	24	28	21
50K−75K	9	14	5
75K+	12	22	2
Don’t know; refused	4	6	2
Number of persons in household				.89
1	29	31	29
2	28	28	29
3	17	19	14
4 or more	26	22	29
SNAP participation (past year)	45	22	64	.0002

SD standard deviation; SNAP Supplemental Nutrition Assistance Program.

^aComparisons between food secure and food insecure with general linear model (bold type of p-values <.05).

^bComparisons between food secure and food insecure with Chi-square or Fisher exact tests (bold type of p-values <.05).

RESULTS

Descriptive statistics and bivariate associations between demographics and food security status are in Table 1. The majority of participants were female (56%), did not attend college (53%), were employed (63%), and had a household income under $25,000 (51%). Just over half of participants (54%) experienced food insecurity in the past year, while less than half participated in SNAP in the past year (45%). Bivariate analysis between food security status and demographics show that food-insecure persons were significantly more likely to be unemployed, have a lower income, and to have received SNAP benefits.

Unadjusted and adjusted regression results are presented in Table 2. In both unadjusted and adjusted models, food-insecure participants, compared to food-secure participants, shopped significantly more often (times/month) at convenience and small grocery stores (18.3 vs. 7.3) and pharmacies (5.2 vs. 2.5). Food-insecure participants also reported shopping more often at dollar stores (10.3 vs. 3.6), though these associations did not reach statistical significance (unadjusted p = .06; adjusted p = .15). Food-insecure persons lived closer than food-secure persons to the store where they bought most of their food and to a small store that they visited in last week, but these differences were also not statistically significant in unadjusted or adjusted models.

Table 2.

Associations between food/beverage shopping-related outcomes and food security status

			Unadjusted		Adjusteda
	Food secure	Food insecure		Model 1	Model 2	Model 3
Outcomes	Descriptive means			Estimate (SE)b
Number of times in past month purchased food or drinks at:
Convenience and small grocery stores, markets, and gas stations	7.3	18.3	11.0 (4.4)*	12.3 (5.0)*	11.7 (5.5)*	11.7 (5.7)*
Dollar stores	3.6	10.3	6.8 (3.6)	5.2 (3.9)	6.1 (4.3)	6.5 (4.4)
Pharmacies	2.5	5.2	2.7 (1.4)*	2.5 (1.4)	2.2 (1.6)	3.1 (1.5)*
Large storesc	13.4	11.1	−2.3 (3.4)	−5.9 (3.7)	−4.1 (4.0)	−4.5 (4.1)
Distance from home (kilometers):
Store where buy most of food for household	5.0	3.9	−1.1 (0.9)	−1.4 (1.0)	−1.3 (1.1)	−1.3 (1.1)
Small store visited in past week for food/beveragesc	3.0	1.5	−1.5 (1.1)	−1.3 (1.3)	−0.8 (1.4)	−0.6 (1.4)
Spending in dollars in past month for food or drinks at:
Convenience and small grocery stores, markets, and gas stations	$65.31	$53.62	−11.7 (13.3)	−1.7 (13.5)	−1.6 (14.8)	−4.1 (15.1)
Dollar stores	$16.69	$32.69	16.0 (6.4)*	9.0 (6.4)	9.1 (6.8)	9.4 (7.0)
Pharmacies	$28.53	$23.36	−5.2 (10.5)	3.0 (9.3)	9.1 (10.1)	11.5 (10.2)
Large storesc	$209.03	$158.93	−50.1 (33.4)	−41.5 (36.9)	−31.7 (40.5)	−30.3 (41.7)
Home Food Inventory (HFI)d score
Fruit (0–26 scale)	3.8	2.6	−1.2 (0.6)*	−0.9 (0.6)	−0.9 (0.7)	−0.7 (0.7)
Vegetables (0–20 scale)	6.1	5.6	−0.5 (0.7)	−0.2 (0.7)	−0.7 (0.8)	−0.6 (0.8)
Obesogenicity scale (1–71)d	17.7	13.6	−4.1 (1.8)*	−4.3 (1.9)*	−2.8 (2.1)	−2.7 (2.1)

^aModel 1: adjusted for age, sex, race, education, employment, household size; Model 2: added household income; Model 3: added Supplemental Nutrition Assistance Program status.

^b  SE standard error; estimates represent how the food-insecure group compared to food-secure group.

^cLarge stores: supermarkets or other large grocery stores (e.g., Target and Costco); small stores: corner, convenience, gas stations, dollar, and pharmacies.

^dFruit and vegetable scales include canned, fresh, and frozen varieties; obesogenicity scale includes regular-fat versions of dairy, frozen desserts, prepared desserts, savory snacks, and added fats; regular-sugar beverages; processed meat; high-fat quick, microwavable foods; candy; and access to unhealthy foods in refrigerator and kitchen.

*Significant p ≤ .05.

In comparisons between food security status and the amount of money spent at various types of stores, the only significant association was among dollar stores (in unadjusted model), with food-insecure persons spending twice as much money in the past month than food-secure persons ($32.69 vs. $16.69; p = .01). For the other store types, persons experiencing food insecurity spent less than food-secure persons, though these differences were not statistically significant in our data.

Finally, the HFI results showed that food-insecure persons had less availability of all types of food availability examined. Differences were significant for both fruit and obesogenic food scales in unadjusted (p = .03) but not fully adjusted models.

Discussion

In this exploratory study, we examined how food insecurity was associated with shopping behaviors and home food environments among frequent small food store shoppers in an urban setting. We build on the growing body of literature in this area that has largely yielded mixed results. We used a sample that was racially and economically diverse and balanced by gender, and we identified the specific type and location of food stores utilized for food/beverage acquisition.

We found that, compared to food-secure persons, persons experiencing food insecurity shopped more frequently at small stores (convenience/small grocery stores, as well as pharmacies), on average, shopping 18 times per month (more than four times per week) at small stores. This was evident in unadjusted models, as well as in models adjusting for household income and whether they received SNAP benefits. This finding is in agreement with findings from several other populations and patient groups [10–12] and points to the important influence of community food environments on those experiencing food insecurity. It is also of note that this result, food-insecure (vs. secure) persons shopping more often at small food stores, was found in our sample, which was comprised of shoppers who were recruited to participate because they reported frequently shopping at small food stores.

Historically, small food stores have not been reliable sources of healthful foods but instead consistently stocked an abundance of unhealthy food [14, 15]. Given that persons experiencing food insecurity tend to have suboptimal dietary outcomes [2–4] and shop frequently at small food stores, interventions aimed at requiring, incentivizing, and/or supporting small stores to stock a minimum level of healthful products, like fresh produce and whole grains, could be an important step to improving access to healthy food, particularly for persons experiencing food insecurity [31]. To offer these foods at affordable prices may require promotions and discounts, possibly with public and/or private subsidies, such as a price discount for healthier versions of products within a category (whole grain vs. white bread) as implemented in the Baltimore Healthy Stores and Healthy Foods Hawaii programs, for example [31]. Placement of healthy foods in prominent locations, such as check-out areas and aisle end caps, similar to several small food store initiatives (e.g., Romano’s Grocery Store Renovation; Healthy Eating, Active Communities [31, 32]), could bolster pricing interventions. In addition, given that a vast majority of small stores accept SNAP benefits [33, 34] and SNAP was designed to benefit persons experiencing or at risk for food insecurity (https://www.fns.usda.gov/snap/supplemental-nutrition-assistance-program), policies requiring retailer stocking standards may be important to maintain for persons using SNAP benefits at these small food stores. For example, the Minneapolis Staple Foods Ordinance (which required food retailers to stock minimum quantities and varieties of 10 product categories, including fruits, vegetables, and whole grains) applied to small food stores and other retailers that accepted federal food assistance programs, like SNAP, under the rationale that these stores are frequently used to purchase groceries among people receiving benefits [35].

Our study adds to the current literature by including data on the amount of money spent on food/beverages at different types of stores and whether this varies by food security status. Food-insecure persons, compared to food-secure persons, reported spending twice as much money on food/beverages in the past month at dollar stores, but differences were not statistically significant in adjusted models for this or other spending outcomes. The one report that examined food-at-home spending by food security status found that spending was lower among food-insecure household but data on spending by store type was not included [36]. Based on our findings and the lack of comparable studies, it is difficult to draw clear conclusions about how the amount of money spent at various types of stores is associated with food security status, independent of income and/or SNAP participation. Future studies with larger samples may be able to more precisely examine these relationships.

We also assessed how objectively measured home food availability varied by food-insecurity status. To our knowledge, only one study has examined the association between food insecurity and home food availability using a validated assessment tool [27]. The few other studies examining this issue to date have relied on self-report and nonvalidated measures [24–26]. We found that households experiencing food insecurity had fewer obesogenic food/beverages than food-secure households but no differences were found for fruits and vegetables in the home after adjustment. These findings are in agreement with a study that used the same validated HFI tool [27], but other studies found mixed results [24–26], making it challenging to provide clear recommendations. A lack of funds may translate to purchasing small amounts of food for immediate or short-term consumption rather than the ability to stock a kitchen with food, healthy or otherwise, or to plan meals for a week or more. More research using validated tools may help to better understand how food security status affects home food environments and how retail food environments may be adapted to allow food-insecure households to expand and improve their home food environments both in terms of food security and healthfulness.

We assessed home-to-store distances for the store usually used for household food shopping and a small food store visited in the past month but found no differences based on food security status. Perhaps, other factors that were not measured in this study, such as access to vehicle or other reliable transportation, may be important in understanding how food insecurity is related to the distances that persons travel to shop for groceries for their households. Given that these transportation options play a deterministic role in how much food can be feasibly transported from stores back to homes, it may be useful to consider these as additional confounders and/or moderators in future research.

Results of this exploratory study should be considered in light of several limitations. First, our sample was limited in size and was drawn from one metropolitan area. This limits generalizability, particularly, to other types of geographic areas and also limited some analytic options, such as distinguishing low versus very low food security status as done in some larger studies (e.g., [11]). Our sample, however, was racially diverse with varying income levels and gender balance, which adds to its generalizability and is a strength compared to other investigations. Although we collected a fair amount of demographic data on the participants, additional information potentially relevant to food insecurity and shopping behaviors, such as transportation options, single- versus dual-parent households, and access to food preparation equipment, could have provided additional insights. An important strength of the paper was recruiting participants from a group of frequent shoppers at small food stores, allowing us to better understand the food-insecurity experience among this unique segment of the population about which little is known. Our results can lend insight into how to create more healthful food environments for persons, both food insecure and secure, who frequently utilize small stores.

Further research into food procurement behaviors among food-insecure persons is needed, taking into consideration the dynamic relationships between income level, SNAP benefits, and experiencing food insecurity. Longitudinal investigations to assess how changes in food security status influence shopping behaviors are a natural next step and could help guide modifications of the food environment to reduce food insecurity.

TRANSLATIONAL IMPLICATIONS

Our study is one of the few to date to examine how food-shopping behaviors and home food environments differ based on food security status independent of other economic confounders. This area of research is important to inform potential interventions, including policies, for improving access to food, particularly healthful options, among food-insecure populations. Small food stores may be important points of intervention for encouraging healthful affordable food choices among food-insecure persons. Our findings lend insight into persons who shop at these small stores, particularly frequent shoppers. Previous studies that rely on national data sets may not be as well suited to understand this population due to small stores often being underrepresented (e.g., [37]) and the lack of objective observation of home food inventory.

Knowing food-insecure persons shop significantly more often at small food stores than food-secure persons, even in a sample of frequent small food store shoppers, provides further evidence that using policy to improve the healthfulness and affordability of products in small food stores may be most beneficial to address these issues. City and state policies that require or incentivize small food stores to stock more healthful and affordable products is one potential avenue. For instance, the Minneapolis Staple Food Ordinance requires all food stores in the city to stock minimum quantities of staple foods across 10 categories (e.g., fruits/vegetables and whole grains). These types of standards [35] could be incorporated into small store intervention programs, such as healthy corner store programs that provide technical assistance to retailers or healthy corner store certification programs (i.e., where a city or other body offers a certification for stores that demonstrate a defined standard of healthfulness, such as minimum stocking of healthy and staple food) [31]. In addition, strengthening the federal minimum stocking standards for SNAP authorization so that SNAP-certified stores must stock minimum varieties of staple foods would be a very beneficial step to improving available foods in small stores since they make up a majority of SNAP-authorized retailers. SNAP stocking standards were scheduled to be strengthened in 2017 but these changes were disrupted by the incoming administration that year. A recent study has shown that these standards did improve the stocking of healthy and stable foods in small food stores [38]. Resurrecting and/or adapting these proposed changes would be a reasonable next step to improving the community food environment. However, as in all policy approaches, it will be as important to ensure that the implementation matches the intention of the written policy to ensure its intended effects on food-insecure populations.

APPENDIX

	Pearson Correlation Coefficients (p-value)
	Food insecure (1 = yes; 0 = no)	Female (1 = yes; 0 = no)	White (1 = yes; 0 = no)	Some college (1 = yes; 0 = no)	Employed (1 = yes; 0 = no)	Household size 4+ (1 = yes; 0 = no)	Age (continuous)	Income: 25K+ (1 = yes; 0 = no)	SNAP (1 = yes; 0 = no)
Corner/small grocery	.27568	−.07862	.02099	.06741	−.12185	.08545	.0191	−.09829	−.08523
Times/month	(.0146)	(.4938)	(.8553)	(.5576)	(.2911)	(.457)	(.8682)	(.4015)	(.4581)
Dollar stores	.21084	.08579	−.23807	−.21806	−.23505	−.10686	.00898	−.15571	−.19427
Times/month	(.0639)	(.4552)	(.0358)	(.0551)	(.0396)	(.3518)	(.9378)	(.1822)	(.0883)
Pharmacies	.21868	.18008	−.22771	−.30797	.0916	.25957	−.16304	−.11473	.06422
Times/month	(.0544)	(.1146)	(.045)	(.0061)	(.4282)	(.0217)	(.1538)	(.327)	(.5764)
Large stores	−.07719	.09203	−.06668	−.06359	−.0282	−.02493	−.23623	.16476	−.04979
Times/month	(.5018)	(.4229)	(.5619)	(.5802)	(.8077)	(.8285)	(.0373)	(.1578)	(.6651)
Home to usual store	−.14468	−.05185	−.07634	−.19845	.10398	.30252	−.12201	.1035	.0757
km	(.2391)	(.6746)	(.5361)	(.1047)	(.3988)	(.0122)	(.3216)	(.4046)	(.5395)
Home to small store	−.17465	−.18371	−.02575	−.04688	.18371	−.13545	−.0701	.20553	.20594
km	(.1898)	(.1675)	(.8478)	(.7268)	(.1713)	(.3107)	(.6011)	(.1286)	(.1209)
$ spent: small stores	−.10011	−.04201	.11746	.02494	.17895	.20773	−.08441	.20883	.0157
In last month	(.3832)	(.715)	(.3058)	(.8284)	(.1194)	(.068)	(.4625)	(.0722)	(.8915)
$ spent: dollar stores	.27625	.18671	−.37184	−.37835	−.34486	.11173	−.05451	−.14478	−.33507
In last month	(.0144)	(.1017)	(.0008)	(.0006)	(.0021)	(.3301)	(.6355)	(.2152)	(.0027)
$ spent: pharmacies	−.05653	.1765	−.04437	−.14669	.11896	.18682	.00741	.2063	.03111
In last month	(.623)	(.1222)	(.6997)	(.2)	(.3028)	(.1015)	(.9486)	(.0758)	(.7869)
HFI: fruit	−.23969	.05224	.25774	.14598	.02519	.02802	.26085	.1601	.21722
1–26 scale	(.0345)	(.6497)	(.0227)	(.2022)	(.8278)	(.8076)	(.0211)	(.17)	(.0561)
HFI: veg	−.08056	.09179	−.04643	−.28821	−.01146	.21461	.24549	−.03137	−.00673
1–220 scale	(.4832)	(.4241)	(.6865)	(.0105)	(.9212)	(.0592)	(.0303)	(.7893)	(.9534)
HFI: obeseogenicity	−.24845	.03838	.1029	−.02688	.05806	.31812	.00791	.32349	.16283
1–71 scale	(.0283)	(.7386)	(.37)	(.8153)	(.616)	(.0045)	(.9452)	(.0046)	(.1543)

SNAP Supplemental Nutrition Assistance Program.

Compliance with Ethical Standards

Conflicts of Interest: K. Lenk, M. Winkler, C. Caspi, and M. Laska declare that they have no conflicts of interest.

Authors’ Contributions: M. Laska, the principal investigator of the STORE study, supervised all aspects of the work; M. Laska, C. Caspi, and M. Winkler devised the main conceptual ideas for the manuscript and helped shape the analysis. K. Lenk reviewed relevant literature, performed all data analyses and wrote the manuscript. All authors discussed the results, contributed to the interpretation of the results, and reviewed the final manuscript.

Ethical Approval: All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. This article does not contain any studies with animals performed by any of the authors.

Informed Consent: Informed consent was obtained from all individual participants included in the study.