Urban versus suburban perceptions of the neighborhood food environment as correlates of adolescent food purchasing

Abstract

Objective

To assess the relationship between adolescent perception of time to walk to neighborhood food retail outlets and purchasing of sugar sweetened beverages (SSB), fast and convenience food items and test for differences by urban versus suburban environment.

Design

Cross-sectional observational study.

Setting

Twin Cities Metropolitan Area, Minnesota.

Subjects

Adolescents from two studies completed survey-based measures on perceptions of time to walk to food retail outlets from home, purchasing patterns of SSB and fast and convenience store items, perceptions of personal safety and pedestrian infrastructure and demographic characteristics. Descriptive analysis, Spearman correlations and multivariate linear regression, accounting for clustering, were conducted using SAS.

Results

There were 634 adolescents, approximately half male, predominately white, with a middle-class background. Greater perceived time to food outlets were associated with less frequent purchasing of SSB, convenience store foods and fast food items. Multivariate models showed that a perceived shorter walking time (i.e. 1–5 versus 31+ minutes) was significantly associated with more SSB purchasing. SSB purchases were also significantly associated with the number of food outlets within a 10–minute walk (B=0.05, p=0.02).

Conclusions

A reduction in consumption of SSB and other energy dense snacks is an important obesity prevention approach. An approach offering alternatives or reducing exposure in addition to education to alter purchasing habits may contribute to improving dietary habits and reducing the obesity epidemic.

Introduction

Adolescent consumption of energy-dense foods and beverages has increased over time, contributing to the obesity epidemic.¹ Data from Project EAT (Eating Among Teens), a longitudinal, epidemiologic study of a large youth cohort, found that eating fast food more than three times per week increased significantly from 1999 to 2004 among adolescent males (from 24% to 30%) and females (19% to 27%)² and average daily sugar sweetened beverage (SSB) consumption increased from 0.94 servings per day to 1.25 servings per day among males as they transitioned from middle to high school.³ The consumption of fast food, SSB, and convenience products which are frequently energy dense, low nutrient products, as well as away-from-home eating in general, have been identified as prime targets for reducing youth obesity.¹

Neighborhood food environments have been charged with contributing to poor dietary intake, with most of the focus in the literature being on the association between individuals’ dietary intake patterns and objectively measured proximity to area food sources using Geographical Information Systems (GIS). As such, a recent review reported that increased access to supermarkets is associated with healthier food intake among adults and adolescents.^{4, 5} Among youth, several studies have also linked proximity to convenience stores with reduced intake of fruits and vegetables, poor overall dietary intake patterns and risk for obesity.^{4, 6–8} However, the literature on associations between diet quality and proximity to fast food restaurants has been mixed. Some evidence suggests proximity to fast food retailers is associated with lower fruit and vegetable consumption⁶ more high fat food⁸ and SSB consumption among adolescents,^8–10 yet others show few or no associations between proximity or access to fast food outlets and overall dietary intake¹⁰ or weight status.^{11, 12}

An individuals’ perception of time required to travel to food retail is important when considering neighborhood ‘accessibility’ among adolescents. Although the actual distance to a food retail outlet measured by GIS may be relatively close, GIS ‘distance to’ does not necessarily include features such as crossing busy streets or navigating multiple traffic lights as noted in physical activity resources literature.¹³ Adolescent perceptions of time required may be more reflective of his or her likelihood of accessing the local food retail. In addition, perceptions of personal safety and pedestrian infrastructure will also play a role in the perceived ‘accessibility’ of food retail, particularly for adolescents.^{14, 15}

No known research to date has examined youth food purchasing behavior from community sources. This is an important gap in the literature. We hypothesize that as youth may be able to more accurately recall how frequently they purchased something (rather than consumed), this may be more closely linked with the neighborhood food retail environment, and youth food purchasing may be an important point of intervention despite the fact that no known previous scientific research has explored these issues specifically.¹⁶ The association between consumption and proximity of neighborhood food stores may be more closely tied to purchases that are made at food retail, and exclude home, school and vending sources.

Therefore, the purpose of this study was to explore the relationship between adolescent perceptions of proximity (i.e., walking time) to neighborhood food retailers and adolescent SSB, fast food and convenience store purchasing behavior. Our hypothesis was that frequency of fast food and SSB purchases would increase as the perceived walking time to food outlets decrease, after taking into account demographic characteristics and perceptions of personal safety and pedestrian infrastructure. Secondarily, we believed that this relationship may be different among youth living in urban areas (i.e., where population density is generally high, mixed land use is prevalent, and street networks are highly connected) versus suburban areas (i.e., where commercial retailers may be more in residential areas, and proximity may play less of a role in youth food purchases). We hypothesized that because of the traditional design of suburban neighborhoods being car-centric versus pedestrian-centric, there would be little association between perceived time to neighborhood food retailers and purchasing behavior in suburban areas but an inverse association in urban areas.

Methods

Data for this analysis came from two etiologic studies of adolescent obesity conducted in Minneapolis and St. Paul, MN and surrounding suburbs. The first sample was from the [removed for review] study is a 3-year longitudinal study aimed at understanding the social and environmental influences on unhealthy weight gain in adolescents.[removed for review] Youth were recruited from a preexisting cohort (The Minnesota Adolescent Community Cohort (MACC) Tobacco Study),¹⁷ contacting adolescents listed on a permit application listing from the Minnesota Department of Motor Vehicles (DMV), and a convenience sample from the St. Paul-Minneapolis metropolitan area. The MACC cohort was recruited from a random sample of 60 geographical and political units (GPU) (out of 129) developed to provide a sample of the local tobacco control environments in the state of Minnesota. The MACC study initially recruited a total of 3,637 teens divided up into five cohorts based on initial age (i.e. 12, 13, 14, 15, 16) which each cohort having approximately 725 youth. Additional information can be found in Widome, 2007.¹⁷ The convenience sample was recruited by asking enrolled participants to notify friends, classmates and other family to contact study staff to determine eligibility. Twenty-six percent of the adolescents were recruited from the MACC cohort, 49% were recruited from the DMV sample, and 25% were recruited from the convenience sample. Participants in this study (n=332) were adolescents (ages 10.8–17.7 years at baseline) and one parent/guardian living in the catchment area. Data for this analysis were collected in 2007–2008.

The second sample was from the [removed for review] study (n=374). Baseline data were collected on adolescents (ages 11.0–17.6) and one parent/guardian in 2007–2008. The participants were recruited from the membership of [removed for review] health plan within the seven-county metropolitan area of Minneapolis, St. Paul, Minnesota. The recruitment plan was designed to increase racial/ethnic diversity and to sample youth and parents that represented both healthy weight and overweight individuals. The [removed for review] studies collected the same measures on all participants from the same target population. Appending the data from the two studies provided a larger and more diverse sample. Both studies were approved by the University of Minnesota Institutional Review Board.

Measures

Adolescents and one adult caregiver (typically a parent) independently completed written surveys under the direction of trained study staff. The survey asked questions related to demographics and perceptions, attitudes and behaviors related to energy balance.

Food and beverage purchasing behavior was measured using three items. Adolescents were asked how many times in the past month they purchased 1) SSBs (regular soda, sports drinks, sweetened teas, juice drinks, punch or lemonade) at a convenience store, gas station, hardware store or vending machine outside of school; 2) food at a conveniences store, gas station, hardware store or a vending machine outside of school; and 3) food at a restaurant where food is ordered at a counter or at a drive-through window (where there is no waiter/waitress). Nine response categories ranged from ‘never or rarely’ to ‘3 or more times per day’. Response categories were recoded to reflect the number of times per day such purchases was made. Test-retest reliability was conducted during pilot testing with 33 participants for the following psychosocial measures. Test retest reliability was ρ=0.67 for all three purchasing items.

Perceived distance to neighborhood food retailers was adapted from the NEWS survey.¹⁸ Perceived walking time to features was approximated using a series of questions asking the adolescent to estimate how long it would take to get from his/her home to the nearest business or facility (convenience store, supermarket, fast food restaurant, coffee place, non-fast food restaurant, or store (video, clothes, pharmacy and drug store)) if s/he walked. Test-retest values for the NEWS subscale among youth was 0.78¹⁸. The response categories were 1–5 minutes, 6–10 minutes, 11–20 minutes, 21–30 minutes, 31+ minutes and treated as an ordinal variable in the analysis using 31+ minutes as the referent category. An index score to represent variety of food features was calculated by summing the number of neighborhood food features the adolescent could walk to in under 10 minutes.

The Concern of Safety scale ¹⁸ was comprised of 5 items (alpha=0.76) with 4 response categories (strongly disagree-strongly agree) and was used as a covariate. Questions include too much traffic, fumes when I walk, lots of crime, safe to walk or play during the day and night. A higher value in the Concern for Safety scale reflects perceptions of more safety concerns and was added as a covariate to the regression models as a potential confounder.

The ease of mobility scale ¹⁸ was comprised of 5 items (alpha=0.68) with 4 response categories (strongly disagree-strongly agree) and was also used as a covariate. Questions include presence of sidewalks, trails, crosswalks and signals, well lit, walkers can be easily seen. A higher value in the Ease of Mobility scale indicates better ease of mobility and was added as a covariate to the regression models as a potential confounder.

Demographic data related to adolescent gender, age, and race were reported by the adolescent and the parent/guardian participating in the study responded to whether or not the adolescent received free or reduced cost lunch at school.

Suburban versus urban classification of neighborhoods was determined by the adolescents’ residential zip code. Adolescents who resided in zip code that were located within U.S. Census Bureau defined urban area (densely settled population inside boundaries of a large incorporated municipality¹⁹) were classified as urban. Suburb designation was those zip codes surrounding the urban area. There were no rural residential households in the data as all participants resided within the 7-county metropolitan area.

Analysis

Sample characteristics and distribution of analytic variables were summarized and tested for differences by urban versus suburban environment using a chi-square test for categorical variables and a t-test for continuous variables. Spearman correlation coefficients tested the bivariate associations between purchasing behavior and perceptions of time to walk to neighborhood food retailers and concerns of safety and ease of mobility. As youth who were recruited were nested in schools, multilevel crude and adjusted linear regression models, accounting for possible clustering at the school level, examined the cross-sectional relationship between purchasing behavior and perception of walking time to neighborhood food features.

A variable distinguishing study sample was included in both crude and adjusted models to account for potential unmeasured confounding by study sample. Adjusted models also included adolescent gender, age, receiving free or reduced cost lunch, race (white versus other), perceptions of safety and ease of mobility. Crude models were tested for two-way interaction between urban/suburban environment and each retail location with p<0.05 considered a statistically significant interaction. If there was no significant interaction, an indicator variable for urban/suburban was added to the adjusted models. If the interaction was significant, urban versus suburban strata specific crude and adjusted models were calculated. All data processing and statistical models were conducted using SAS v9.1 for Windows.²⁰

Results

The combined sample yielded n=706 participants. The urban/suburban designation was collected at the home level and 72 students were missing data or were missing the urban/suburban designation, resulting in a final analytic sample size of n=634. Descriptive results are shown in Table 1. There were no significant differences in the sample composition of gender by urban/suburban environment, but in suburban areas there were a higher percentage of adolescents who were white (90.0% versus 65.4%, chi-square=53.7, p<0.001) and a lower percentage of adolescents receiving free or reduced cost lunch (7.0% versus 24.1%, chi-square=35.0, p<0.001) compared to urban areas. Those living in urban environments were also younger than those in the suburbs (mean age_u=14.3 versus mean age_s=15.3, t = 4.1, p<0.001).

Table 1.

Descriptive characteristics of the neighborhood food environment by urban and suburban environment, TREC IDEA and ECHO 2007

		Total Sample N=634		Urban N=162		Suburban N=472		Chi-square	p-value
		%		%		%
Male		48.6		53.1		47.0		1.77	0.18
White		83.8		65.4		90.0		53.7	<0.001
Receive free or reduced price lunch		11.4		24.1		7.0		35.0	<0.001
		Mean	Std Dev	Mean	Std Dev	Mean	Std Dev	T statistic	P value
Age		15.0	2.1	14.5	2.1	15.3	1.9	4.1	<0.0001
Time to walk to the nearest(minutes)…		%		%		%		Chi2
convenience store	1–5 min	31.4		45.1		26.7		35.2	<0.0001
	6–10 min	24.9		28.4		23.9
	11–20 min	26.1		21.0		24.5
	21–30 min	8.8		4.3		10.4
	31+ min	8.8		1.2		11.4
supermarket	1–5 min	11.4		15.4		10.2		9.9	0.04
	6–10 min	19.9		17.9		20.6
	11–20 min	23.2		29.0		21.4
	21–30 min	14.8		14.2		15.0
	31+ min	30.8		23.5		32.8
fast food restaurant	1–5 min	14.0		19.1		12.5		14.6	0.01
	6–10 min	18.5		21.0		17.6
	11–20 min	27.8		25.9		28.4
	21–30 min	19.7		22.8		18.4
	31+ min	20.0		11.1		23.1
coffee place	1–5 min	18.5		29.0		15.0		20.5	0.004
	6–10 min	21.0		21.0		21.0
	11–20 min	21.5		22.2		21.4
	21–30 min	17.9		14.8		18.9
	31+ min	21.0		13.0		23.7
non-fast food restaurant	1–5 min	16.1		28.1		12.1		31.6	<0.0001
	6–10 min	21.2		20.6		21.4
	11–20 min	20.9		23.8		20.3
	21–30 min	16.5		13.8		17.4
	31+ min	25.3		13.8		28.8
store†	1–5 min	15.5		24.1		12.5		28.0	<0.0001
	6–10 min	22.6		28.4		20.8
	11–20 min	25.1		23.5		25.9
	21–30 min	18.6		16.7		19.1
	31+ min	18.3		7.4		21.8
# features < 10 min. away (range 0–6)		2.4	2.2	3.0	2.3	2.1	1.9	−4.3	<0.0001
Perceptions of
Safety concerns		8.7	2.5	10.1	2.3	8.3	2.6	−8.3	<0.0001
Ease of mobility		13.9	3.0	15.2	3.1	13.4	2.3	−6.5	<0.0001
Daily Purchasing Behavior (average)
Sugar sweetened beverages		1.0	1.1	1.0	1.7	0.8	1.8	−1.1	0.11
Convenience store purchases		0.7	1.5	1.1	1.1	0.8	2.4	−3.6	0.0003
Fast Food purchases		0.9	1.7	1.1	1.1	0.8	1.2	−0.9	0.39

^†Store category includes video, clothes, pharmacy and drug store.

Thirty-one percent of the sample reported they could walk to a convenience store in 5 minutes or less, as compared to 11% to a supermarket, 14% to a fast food restaurant, 18% to a coffee place, 15% to a non-fast food restaurant and 15% to another type of store. Adolescents reported an average of 2.4 types of neighborhood food retailers within a 10 minute walk from their home. There were several statistically significant differences in perceived walking time between urban compared to suburban adolescents. See Table 1. Adolescents in urban areas reported generally lower walking times to convenience stores (chi-square=31.3, p=<0.001); fast food restaurants (chi-square=9.3, p=0.05); coffee place (chi-square=15.1, p=0.005); non-fast food restaurants (chi-square=21.5, p<0.001); and stores (chi-square=25.8, p<0.001). However, there were no significant differences in walking times to a supermarket between urban and suburban environments.

Spearman correlations (Table 2) showed daily purchasing of SSBs was significantly correlated with proximity to all neighborhood food features. As the length of time to walk to each type of retailer decreased (e.g. supermarket being 21–30 minutes away versus 31+ minutes away), the number of daily SSB purchases increased. As the variety of neighborhood food retailers within a 10 minute walk increased, adolescents’ daily SSB and convenience store food purchases also increased. Daily purchasing of food at a convenience store was also significantly correlated with the distance to convenience stores, non-fast food restaurants and stores. Thus, as the perceived walking time increased (e.g. 1–5 minutes to 6–10 minutes), daily purchasing of food from convenience stores decreased. As perceived walking time increased to the nearest convenience store, supermarket and fast food restaurant, daily fast food purchases decreased.

Table 2.

Correlations between purchasing behavior and features of the neighborhood food environment.

	Daily Sugar Sweetened Beverage Purchases	Daily convenience store purchases	Daily fast food purchases
Time to walk to the nearest(minutes)…
convenience store	−0.20**	−0.13**	−0.10**
supermarket	−0.12**	−0.07	−0.09*
fast food restaurant	−0.17**	−0.06	−0.08*
coffee place	−0.12**	−0.06	0.00
non-fast food restaurant	−0.15**	−0.09*	−0.05
store	−0.14**	−0.08*	−0.05
# features < 10 min. away (range 0–6)	0.17**	0.10**	0.06

^*p<0.05;

^**p<0.01

Table 3 presents the results from the crude and adjusted multi-level models examining associations between food purchasing behavior and perceived time to walk to the nearest neighborhood food retailer. There were no significant interactions by suburban/urban environment in the association between time to walk to neighborhood food retailers and purchasing behavior. Crude models show a stepped decrease in SSB purchasing by shorter versus longer walking time to convenience stores, supermarkets, fast food restaurants, coffee places, non-fast food restaurants and stores. Results were consistent in adjusted models, although slightly attenuated. For example, participants who reported a convenience stores was 1–5 minutes, compared to 31+ minutes away by walking, reported purchasing more SSB daily (β=0.84, p<0.01). There was only a minimal association between fast food purchases and proximity to a supermarket; participants reported more fast food purchases if the walking distance to a supermarket was 1–5 minutes away compared to 31+ minutes (β=0.32, p<0.05). No other associations were significant for convenience store or fast food purchases in adjusted models.

Table 3.

Crude and adjusted multivariate linear regression of purchasing behavior and neighborhood food environment.

B (SE)		Sugar Sweetened Beverage Purchase				Convenience Store Purchase				Fast Food Purchases
		Crude		Adjusted		Crude		Adjusted		Crude		Adjusted
		Coeff	SE	Coeff	SE	Coeff	SE	Coeff	SE	Coeff	SE	Coeff	SE
Time to walk to the nearest…	Minutes
convenience store	1–5 min	1.06**	0.18	0.84**	0.22	0.60*	0.16	0.22	0.18	0.23	0.15	0.15	0.18
	6–10 min	0.79**	0.21	0.64**	0.22	0.30	0.18	0.02	0.19	0.13	0.16	0.08	0.17
	11–20 min	0.60**	0.17	0.59**	0.18	0.22	0.16	0.11	0.18	0.08	0.16	0.08	0.18
	21–30 min	0.25	0.19	0.20	0.21	−0.07	0.20	−0.17	0.21	−0.15	0.17	−0.16	0.23
	31+ min	Ref		Ref		Ref		Ref		Ref		Ref
supermarket	1–5 min	0.57**	0.19	0.49**	0.20	0.22	0.17	0.10	0.16	0.32*	0.15	0.32*	0.15
	6–10 min	0.46**	0.18	0.38*	0.17	0.16	0.15	0.06	0.15	0.16	0.13	0.16	0.13
	11–20 min	0.38*	0.17	0.33*	0.18	0.26	0.16	0.17	0.16	0.19	0.12	0.20	0.12
	21–30 min	0.31	0.17	0.36*	0.17	0.06	0.17	0.09	0.16	0.32*	0.14	0.36*	0.15
	31+ min	Ref		Ref		Ref		Ref		Ref		Ref
fast food restaurant	1–5 min	0.89**	0.18	0.68**	0.19	0.23	0.19	−0.06	0.20	0.29*	0.15	0.22	0.17
	6–10 min	0.65**	0.17	0.48**	0.18	0.07	0.16	−0.18	0.17	0.21	0.13	0.16	0.15
	11–20 min	0.41**	0.14	0.37**	0.14	0.11	0.16	−0.02	0.17	0.10	0.12	0.11	0.14
	21–30 min	0.42*	0.19	0.38*	0.18	−0.03	0.20	−0.12	0.19	0.18	0.13	0.16	0.14
	31+ min	Ref		Ref		Ref		Ref		Ref		Ref
coffee place	1–5 min	0.60**	0.19	0.50**	0.19	0.32	0.17	0.15	0.17	0.01	0.16	−0.01	0.17
	6–10 min	0.35*	0.18	0.34*	0.18	0.10	0.16	0.06	0.16	−0.02	0.15	−0.003	0.15
	11–20 min	0.29	0.19	0.24	0.18	0.04	0.18	−0.03	0.17	0.12	0.15	0.08	0.15
	21–30 min	0.27	0.19	0.20	0.17	0.24	0.18	0.16	0.17	0.18	0.15	0.16	0.15
	31+ min	Ref		Ref		Ref		Ref		Ref		Ref
non-fast food restaurant	1–5 min	0.69**	0.18	0.55**	0.18	0.35*	0.17	0.16	0.17	0.07	0.15	0.02	0.16
	6–10 min	0.52**	0.17	0.40*	0.17	0.23	0.17	0.09	0.17	0.15	0.13	0.12	0.13
	11–20 min	0.49**	0.18	0.42*	0.18	0.10	0.15	0.003	0.16	0.19	0.13	0.18	0.14
	21–30 min	0.28	0.17	0.26	0.15	0.03	0.16	0.02	0.14	0.12	0.13	0.09	0.13
	31+ min	Ref		Ref		Ref		Ref		Ref		Ref
Store	1–5 min	0.57**	0.19	0.39*	0.19	0.27	0.19	0.01	0.18	0.13	0.16	0.10	0.17
	6–10 min	0.41*	0.19	0.32	0.20	0.06	0.16	−0.09	0.18	0.14	0.15	0.15	0.15
	11–20 min	0.33	0.19	0.28	0.19	−0.02	0.18	−0.11	0.17	0.13	0.15	0.13	0.15
	21–30 min	0.03	0.20	−0.03	0.20	−0.20	0.19	−0.31	0.18	0.15	0.14	0.16	0.15
	31+ min	Ref		Ref		Ref		Ref		Ref		Ref
# features < 10 min. away		0.07**	0.02	0.05*	0.02	0.03	0.02	0.01	0.02	0.01	0.02	0.005	0.02

^† Crude model includes adjustment for study and clustering at school. Adjusted includes cluster at school, study, gender, age, receiving free and reduced cost lunch, race (white versus other), perceptions of safety, ease of mobility, and urban versus suburban.

^*p<0.05;

^**p<0.01

Discussion

These cross-sectional analyses support our hypothesis that closer perceived proximity (or travel time on foot) to food retail is related to increased SSB purchasing among adolescents. However, convenience store and fast food restaurant purchasing was minimally associated with perceived walking time to food retail after adjusting for covariates, such as pedestrian infrastructure, perceptions of safety and sociodemographic factors. Contrary to our hypothesis, there were no significant differences in these relationships in suburban versus urban environments. It is possible that there is a linear relationship between perceptions of proximity to food retailers and purchasing behaviors that does not differ (in magnitude or direction) between suburban versus urban areas. The differences observed in the sample description stratified by suburban versus urban environment (Table 1) may in fact simply reflect perception of walking distance, e.g. further distance to convenience stores leads to less purchasing, rather than differential effects related to suburban versus urban environment.

Our findings were consistent with a recent study by Laska, et al (2010) that explored the relationship between objective measures of neighborhood food environment using GIS-generated buffers and dietary intake in the [removed for review] Study.¹⁰ The authors found few associations between fast food purchasing, dietary intake and body composition and objectively measured neighborhood food environments, with the exception of a relatively stable association between density and distance to food retail and SSB intake. The current study adds another layer of understanding to this work. Our findings indicate that youth purchasing of SSBs in community settings was more frequent when the perceived walking distance was closer compared to further. In fact, more food retail stores available within a 10-minute walk from home was associated with more SSB purchasing behavior.

The consistency of the relationship between SSB and perceived walking distance to food retail may be related to widespread access. Energy-dense foods are widely available in communities, not only restricted to food stores and restaurants. Farley, et al (2009) assessed 1,082 retail stores that were not primarily food venues, such as pharmacies, gas stations or furniture stores and found snack foods were available at 41% of non-food stores, with the most common items being candy (33%), sweetened beverages (20%) and salty snacks (17%)²¹ The universal availability of energy-dense foods contributes to impulse purchasing and consumption,²¹ in particular SSBs as supported here and in previous research.¹⁰

This study highlighted not only neighborhood exposure to food retail, but also youth food purchasing. Because our youth food purchasing survey questions were designed to intentionally exclude food obtained from home and items purchased from school, these findings offer a clearer link to community-level food retail as an exposure and potential for intervention. Secondarily, the focus on adolescent perceptions of walking distances to food retail is relatively new to the field as most of the published literature that has focused on perceptions of distances has been related to physical activity outcomes.^{22, 23}

Many recent neighborhood food environment studies have made the point that supermarkets provide healthy options for food purchasing and should be encouraged in neighborhoods.⁴ However, in this study, we also observed a positive relationship between perceived proximity to supermarkets and consumption of less healthy options, such as SSB and fast food, among our adolescent sample. Supermarkets and grocery stores sell many healthy foods; however they sell many unhealthy foods as well. This point is often overlooked in the current literature on food environment. In addition, supermarkets may be located in close proximity to other food venues, such as fast food restaurants.

While this study added to the current literature, there were several limitations. The data were cross-sectional, restricting causal inference, and the data on purchasing was not linked to an actual food retail location. It is possible that youth food purchases were made at convenience stores and other food retail locations away from the neighborhoods, including near their schools, in which these youths lived. In a recent study, Laska et al found that among the young adult population (18–23 years of age), most food purchasing behavior occurs outside of traditionally-defined GIS-defined buffers of two miles or less.²⁴

CONCLUSIONS

Research has shown that youth who shop at convenience stores frequently purchase energy dense, low-nutrient foods and beverages that contain approximately 350 calories per purchase, on average.²⁵ Furthermore, children and adolescents today derive 10–15% of their daily calories from SSB, and do not appear to adjust their total dietary intake to compensate for the calories in these beverages.^{26, 27} Given these issues and the findings from our study, a reduction in consumption of SSB and other energy-dense snacks is an important dietary intervention to help stem the obesity epidemic. This suggests that an approach focused on education of adolescents of the caloric content of SSB and an environmental intervention, such as reducing the ubiquity of SSB in the environment, enhancing product placement strategies of healthy food at food retail locations or improving nutrition labeling as a means to reduce exposure and alter purchasing habits may assist in improving dietary patterns and subsequently reducing the obesity epidemic.