Neighborhood Economic Changes After the Great Recession and Home Food Environments

Abstract

Objective.

Home food environments are important contexts for children and their food intake. It is unknown whether neighborhood economic context plays a role in explaining the association between a national economic crisis and children’s home food environments. This study attempts to investigate neighborhood economic changes after the Great Recession and their associations with home food environments.

Method.

Using data from the Geographic Research on Wellbeing survey (2012–2013), we conducted a series of logistic regression analyses to examine the association between neighborhood changes after the Great Recession and home food environments.

Results.

Findings showed that neighborhood economic changes after the Great Recession were concentrated in poor neighborhoods. In addition, our findings demonstrated that poor families residing in neighborhoods severely affected by the Great Recession were vulnerable to less availability of fruits and vegetables in the home after the Great Recession.

Discussion.

Findings imply that public health interventions aiming to improve home food environments should include strategies at the national and neighborhood levels as well as the family level. A priority population for public health interventions should be poor families living in disadvantaged neighborhoods.

Home food environments are important contexts for children and their food intake because children of all age groups consume two thirds of daily energy at home (Poti & Popkin, 2011). Children in healthy home food environments–such as fruits and vegetables available at home, and having family meals at home, are likely to consume more fruits and vegetables and less energy-dense and nutrient-poor snack foods and beverages (Arcan et al., 2007; Blanchette & Brug, 2005; Ding et al., 2012; Golan & Weizman, 2001; Hanson, Neumark-Sztainer, Eisenberg, Story, & Wall, 2005; J. A. Jackson et al., 2017; Larson, Story, Wall, & Neumark-Sztainer, 2006; Neumark-Sztainer, Wall, Perry, & Story, 2003; Wyse, Campbell, Nathan, & Wolfenden, 2011). Home food environments also explained socioeconomic differences in consumption of healthy and unhealthy foods among children (Ranjit et al., 2015). A study found that significant differences in healthy diets across parental education levels disappeared after home food environment measures were incorporated in the model (Ranjit et al., 2015). Further, unhealthy home food environments could be a threat to behavioral, emotional, cognitive, and academic development of children in all stages of childhood, such as self-control, hyperactivity, antisocial behavior, and academic achievement (Faught, Williams, Willows, Asbridge, & Veugelers, 2017; D. B. Jackson, 2016; D. B. Jackson, Newsome, Vaughn, & Johnson, 2018; D. B. Jackson & Vaughn, 2017; Johnson & Markowitz, 2018; Shankar, Chung, & Frank, 2017).

The ecological model of the home food environment (Rosenkranz & Dzewaltowski, 2008) posits that a national economic crisis influences families’ home food environments by altering neighborhood socioeconomic conditions as well as availability and financial costs of foods brought to market (Rosenkranz & Dzewaltowski, 2008). Likewise, a model of large-scale economic change and child development (Weiland & Yoshikawa, 2012) suggests a neighborhood-level mechanism through which a national economic crisis can influence family environments and their children. A national economic crisis can decrease neighborhood-level aggregate employment opportunities and wages and increase residential instability, all of which are associated with fewer community resources, lower social cohesion, and higher crime rates (Weiland & Yoshikawa, 2012), which affect home food environments (Ding et al., 2012; Sampson, Morenoff, & Earls, 1999; Shier, Nicosia, & Datar, 2016) and home food security (Dean & Sharkey, 2011; D. B. Jackson, Johnson, Vaughn, & Hinton, 2019; Nebbitt, Lombe, Chu, Sinha, & Tirmazi, 2016). However, it is not empirically known whether neighborhood economic context plays a role in explaining the association between a national economic crisis and children’s home food environments. To address this limitation, this study attempts to investigate neighborhood economic changes after the Great Recession (as a national economic crisis) and their associations with home food environments in a large, statewide representative, and ethnically diverse sample.

The Great Recession and Neighborhoods

The Great Recession, the most recent, longest, and devastating economic downturn since the Great Depression in 1930s, disproportionally affected socioeconomically disadvantaged or minority neighborhoods because their residents were more likely to hold unstable jobs sensitive to macroeconomic conditions (Allen, 2013; Delmelle & Thill, 2014; Hyra & Rugh, 2016; Ong et al., 2003; A. Owens & Sampson, 2013; Solari, 2012; Williams, Galster, & Verma, 2013). Another reason for the concentrated influence is a high percentage of home mortgages (particularly subprime loans) in many low-income and minority neighborhoods that experienced gentrification, driven in part by the real-estate bubble in the early- to mid-2000s (Hyra & Rugh, 2016). From 2000 to 2010, residents’ median household income in economically advantaged neighborhoods increased while it dropped in disadvantaged neighborhoods (Solari, 2012). The proportion of unemployed residents disproportionally increased in disadvantaged neighborhoods compared with advantaged neighborhoods (A. Owens & Sampson, 2013).

Neighborhoods which were more affected by the recession could have less tax revenue for maintaining health-promoting environments (Chernick, Langley, & Reschovsky, 2011; Weiland & Yoshikawa, 2012) and experience lower quality social environments (Sampson, Raudenbush, & Earls, 1997), which negatively affects home food environments. One study that supports this idea reported that people in more economically depressed counties turned inward, that is, they focused on taking care of themselves or their families rather than others in these communities, which resulted in lower social cohesion and trust (L. A. Owens & Cook, 2013).

Recessions and Home Food Environments

Concerning home food environments during the Great Recession, prior literature showed that families reduced food expenditures by purchasing more carbohydrates-based foods and less fruits and vegetables during the Great Recession—when food prices rose sharply and household disposable income decreased (Bruening, MacLehose, Loth, Story, & Neumark-Sztainer, 2012; Griffith, O’Connell, & Smith, 2013; Hall & Perry, 2013). The rate of food insecurity increased from 11.4% to 15.4% on average and from 20.5% to 25.9% among low- and middle-income households (Anderson, Butcher, Hoynes, & Schanzenbach, 2014). Moreover, the rate of food insecurity was higher by 1.6% for households living in states where the median duration of unemployment was longest (Anderson et al., 2014). All aforementioned studies indicated vulnerability in home food environments from the Great Recession; however, none has compared economic conditions after the Great Recession with those before the Great Recession as a correlate of home food environments in a multivariate analysis. Also, it is unknown whether neighborhood-level changes acted as a mechanism by which the Great Recession may have influenced home food environments independent of family economic status.

The Present Study

This study examined (1) whether any adverse changes in neighborhood economic context between 2000 and 2009–2013 were more pronounced in poor or minority neighborhoods compared with more affluent or non-Hispanic White neighborhoods and (2) whether changes in neighborhood economic context between 2000 and 2009–2013 were associated with home food environments.

The impact of the Great Recession on individuals may differ by family economic status. Clearly, adversities resulting from the Great Recession have been concentrated among low-income workers (Sum & Khatiwada, 2010). Also, poor households are more vulnerable to a given proportional income loss than households whose income levels were high or much above “subsistence” level (Ferreira & Schady, 2009). This hypothesis corroborates findings that the Great Depression aggravated the health of the working class who suffered more from economic changes than those in a higher social class (Doyal & Pennell, 1797). To capture the differential impacts of the Great Recession by family economic status, we examined research questions by family poverty status.

Method

Data Sources

The Geographic Research on Wellbeing (GROW) study is a follow-up survey of postpartum women who participated in the statewide-representative Maternal and Infant Health Assessment baseline survey in California during 2003–2007. Home addresses in the GROW study were geocoded to the census tract level with a geocoding accuracy of 97%. GROW data, collected in 2012–2013, reflect home food environments lagging the Great Recession by several years (Modrek, Stuckler, McKee, Cullen, & Basu, 2013). To obtain measures of neighborhood economic context, we used the Neighborhood Change Database for the year 2000 and the American Community Survey for years 2009–2013. The Neighborhood Change Database corresponds to 2010 census tract boundaries so that the 2000 data represent the exact same geographic boundaries as the American Community Survey 2009–2013 does (GeoLytics, 2013). Neighborhood data sets were merged with GROW data based on residential address at the time of the GROW survey. All procedures involving human participants were approved by the institutional review boards at each institution.

We included households living in the same census tract both at the time of Maternal and Infant Health Assessment baseline survey and GROW. The analytic data set included households (1) whose census tracts had available information on the independent variables (deleting 79 cases) and (2) who provided data on the dependent variables (deleting 7 cases). We excluded households that had missing data on covariates (n = 79) and households that did not live with a child (n = 37). We included Asian/Pacific Islander, Black, Hispanic, and White mothers but removed American Indian/Alaska Native (n = 3) due to small sample size. We also excluded Asian/Pacific Islander families with income below 200% of the federal poverty level due to the small sample size in the analytic model stratified by family poverty (n = 15). The final analytic sample is thus 1,359 households.

Measures

Dependent Variables.

The dependent variables were mother-reported availability of vegetables in the home (excluding potatoes; very often = 1; otherwise = 0), availability of fruits in the home (very often = 1; otherwise = 0), and frequency of cooking dinner at home (excluding precooked food; 5–7 days a week = 1; otherwise = 0).

Independent Variables.

The independent variables were changes in three neighborhood-level indicators before/after the Great Recession: (1) median household income, (2) proportions of vacant housing units, and (3) median housing value. Neighborhoods were defined as census tracts containing approximately 4,000 to 8,000 individuals (Diez-Roux et al., 1997).

A difference in absolute value of the neighborhood economic indicators between years 2000 and 2009–2013 was estimated for each indicator. For example, using median household income, we subtracted the value of median household income in years 2009–2013 by that in year 2000 and then categorized neighborhoods into four groups: (1) decreased; (2) increased by $0 to 9,999; (3) increased by $10,000 to 14,999; or (4) increased by $15,000 or more. The cutoffs were determined by considering quartiles of the variable. Similar methods were used to measure differences in the other two variables.

Individual-Level Variables.

Mothers reported the previous calendar year’s income from all sources, and the reported values were dichotomized into two groups using 200% of the federal poverty threshold (FPT) as the cut point: (1) families with income below 200% of the FPT (hereafter, poor families) and (2) families with income 200% of the FPT or above (hereafter, nonpoor families). In addition, mother’s age, race/ethnicity, marital status, education level, and employment status, and the number of children living with the mother were included as covariates.

Data Analysis

First, bivariate analyses were conducted to examine neighborhood economic changes between years 2000 and 2009–2013 by neighborhood-level non-Hispanic White and poverty rates before the recession. Second, another set of bivariate analyses was conducted to examine home food environments by individual-level and neighborhood-level characteristics separately for poor and nonpoor families by using GROW study data linked to the neighborhood data. Third, logistic regression models were estimated to obtain odds ratios for each dependent variable by poor and nonpoor families: a “sociodemographic” model (Model 1); “neighborhood economic change” models (Models 2–4), which added variables of neighborhood economic change one at a time to the sociodemographic model; and a fully adjusted model (Model 5), which added all three variables of neighborhood economic change to the sociodemographic model. Since 90% of census tracts included only one or two GROW respondents, this study did not use formal multilevel modeling. Analyses were weighted and took into account the complex sample design using SAS 9.4.

Results

Neighborhood Economic Changes After the Great Recession

Deteriorating neighborhood changes after the recession were concentrated in previously White or high-poverty neighborhoods (see Figure 1). For example, neighborhoods with the lowest quartile of non-Hispanic White rates had a lower increase in median household income than those in its third or fourth quartile ($8,492 vs. $10,835–10,921). Similarly, median household income increased by approximately $8,430 in previously high-poverty neighborhoods (vs. $13,516 increase in low-poverty neighborhoods).

Figure 1.

Neighborhood economic changes between 2000 and 2009–2013 by neighborhood-level non-Hispanic White rates and poverty rates before the Great Recession.

Note. Census tracts in the analytic sample were analyzed; all chi-square values are statistically significant.

Sample Characteristics

As Table 1 shows, the majority of mothers included in this study were married or living with a partner (85%), lived with two or more children (90%), and were high school graduates (80%). Median household income, proportions of vacant housing units, and median housing values increased in most neighborhoods where the participating mothers resided. Availability of vegetables and fruits generally increased with mothers’ age and was highest for non-Hispanic Whites (poor families) or foreign-born Latinas (nonpoor families), households with more children, and higher socioeconomic status, although the patterns varied somewhat for poor versus nonpoor families. The percent of households cooking dinner at home frequently was lowest for non-Hispanic Black mothers, unmarried mothers, fewer children living in the household, and higher socioeconomic status. Patterns in home food environments were less clear for the neighborhood economic improvement variables.

Table 1.

Individual- and Neighborhood-Level Characteristics of Respondents (Weighted), Geographic Research on Wellbeing (GROW) Study, California, United States, 2012–2013, N = 1,359

		Home food environments (%)
		Poor families (n = 575)			Nonpoor families (n = 784)
Characteristic	Total (%)	Vegetable in home	Fruit in home	Cooking dinner at home	Vegetable in home	Fruit in home	Cooking dinner at home
Total	100.0	78.5	81.0	83.8	91.1	93.8	62.9
Mother’s age (years)
20–29	16.1	73.2	77.6	84.9	94.8	91.6	68.7
30–39	44.0	80.2	81.5	84.3	92.5	93.1	64.9
40+	39.9	80.8	83.5	81.6	89.6	94.5	60.7
Mother’s race/ethnicity
Asian/Pacific Islander	14.1	—	—	—	90.3	91.3	67.9
Black	6.0	74.5	81.6	63.4	87.5	87.3	54.3
Latina, foreign-born	37.9	78.1	79.5	89.1	95.4	98.4	70.7
Latina, U.S.-born	16.0	77.4	80.8	76.7	85.4	90.2	53.4
White, non-Hispanic	26.0	87.2	92.4	72.3	93.1	96.4	62.4
Mother’s marital status
Married or living together	85.3	79.1	82.2	87.2	91.3	98.2	64.3
Unmarried	14.7	76.7	77.0	72.6	86.6	93.5	38.4
Number of children in household
1	10.5	66.0	70.6	76.8	87.9	93.7	53.5
2	39.5	75.0	78.4	75.4	92.3	95.5	61.5
3+	50.0	81.3	83.2	88.0	90.6	91.2	68.6
Mother’s education
Less than high school	20.0	75.4	76.3	87.8	85.2	84.0	74.0
High school graduate/GED/some college	43.3	81.5	84.5	81.5	86.4	89.9	65.3
College graduate or above	36.7	70.6	77.2	79.3	93.4	95.8	61.4
Family income (% of federal poverty level)
≤100%	29.5	79.4	80.6	87.3	—	—	—
101–200%	20.4	77.2	81.5	78.6	—	—	—
201–300%	9.3	—	—	—	88.7	91.2	68.4
301–400%	7.8	—	—	—	89.5	93.5	64.4
>400%	33.0	—	—	—	92.1	94.5	60.9
Mother’s employment status
Did not work	46.0	78.3	80.1	90.0	89.6	92.5	71.6
Part-time work	10.3	75.8	83.9	82.5	93.1	95.8	63.2
Full-time work	43.7	80.0	81.8	71.9	91.5	94.1	57.9
Difference in median household income
<$0	15.2	65.5	69.4	81.5	93.6	98.5	68.0
$0-$100,000	36.6	78.4	80.4	85.7	89.7	95.4	64.4
$100,000-$149,999	19.9	83.7	86.4	86.3	91.6	93.9	63.0
$150,000+	28.3	83.3	85.2	78.3	90.7	90.4	59.5
Difference in % of vacant housing units
<0.00%	29.2	76.2	77.2	85.5	91.3	92.8	61.6
0.00–2.99%	36.7	79.9	81.8	84.5	93.3	94.1	64.4
3.00–4.99%	16.6	83.7	86.7	86.0	88.2	92.5	63.9
5.00%+	17.5	74.6	79.8	77.5	88.5	95.9	60.8
Difference in median housing values
<$100,000	18.0	78.1	77.0	88.9	91.3	90.2	65.8
$100,000-$149,999	24.0	79.8	83.9	84.8	89.7	95.1	64.4
$150,000-$249,999	33.5	76.1	79.0	80.7	90.3	92.6	62.9
$250,000+	24.5	83.5	87.9	78.8	92.3	95.2	61.2

Neighborhood Economic Changes and Home Food Environments

As shown in Model 2, Table 2, living in neighborhoods that experienced an increase in median household income by less than $10,000, $10,000 to $14,999, or $15,000 or more was associated with increased odds of having vegetables in the home very often among poor families (compared with those in neighborhoods that experienced a decrease in median household income) (odds ratios [ORs] = 2.15, 3.23, and 3.10, respectively). Median household income remained significant when including all the three variables of neighborhood economic change in the analysis (see Model 5). Differences in the proportions of vacant housing units and median housing values were not significantly associated with having vegetables in the home very often.

Table 2.

Logistic Regression Analysis Assessing Associations Between Changes in Neighborhood Economic Indicators After the Great Recession and Availability of Vegetables in the Home Among Poor Families, Geographic Research on Wellbeing (GROW) Study, California, United States, 2012–2013, n = 575.

	Model 1		Model 2		Model 3		Model 4		Model 5
Characteristic	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI
Mother’s age	1.03	[0.99, 1.07]	1.04	[1.00, 1.07]	1.03	[0.99, 1.07]	1.03	[0.99, 1.07]	1.04	[0.99, 1.08]
Mother’s race/ethnicity
Black	0.48	[0.17, 1.36]	0.52	[0.17, 1.57]	0.47	[0.17, 1.32]	0.50	[0.18, 1.42]	0.51	[0.17, 1.56]
Latina, foreign-born	0.53	[0.22, 1.23]	0.49	[0.20, 1.20]	0.50	[0.21, 1.16]	0.54	[0.23, 1.28]	0.48	[0.19, 1.20]
Latina, U.S.-born	0.62	[0.23, 1.66]	0.57	[0.20, 1.59]	0.61	[0.23, 1.63]	0.62	[0.23, 1.67]	0.56	[0.20, 1.60]
White, non-Hispanic	1.00		1.00		1.00		1.00		1.00
Mother’s marital status
Married or living together	1.05	[0.58, 1.89]	0.90	[0.50, 1.60]	1.01	[0.56, 1.81]	1.05	[0.59, 1.88]	0.89	[0.50, 1.58]
Unmarried	1.00		1.00		1.00
Number of children in household
1	1.00		1.00		1.00		1.00		1.00
2	1.49	[0.63, 3.52]	1.36	[0.57, 3.27]	1.60	[0.67, 3.80]	1.55	[0.65, 3.68]	1.46	[0.59, 3.62]
3+	2.15	[0.95, 4.82]	2.15	[0.94. 4.91	2.28	[0.99, 5.24]	2.18	[0.96, 4.95]	2.23	[0.94, 5.33]
Mother’s education
Less than high school	1.00		1.00		1.00
High-school graduate/GED	1.50	[0.92, 2.47]	1.47	[0.89, 2.41]	1.46	[0.89, 2.41]	1.49	[0.91, 2.45]	1.43	[0.87, 2.35]
College or above	0.88	[0.37, 2.05]	0.80	[0.34, 1.92]	0.87	[0.36, 2.11]	0.87	[0.36, 2.06]	0.83	[0.33, 2.08]
Mother’s employment status
Did not work	1.00		1.00		1.00		1.00		1.00
Part-time work	0.89	[0.40, 1.97]	0.89	[0.40, 2.00]	0.83	[0.38, 1.84]	0.88	[0.40, 1.96]	0.85	[0.38, 1.90]
Full-time work	1.13	[0.69, 1.85]	1.13	[0.69, 1.83]	1.10	[0.67, 1.82]	1.11	[0.68, 1.83]	1.11	[0.68, 1.83]
Difference in median household income
<$0			1.00						1.00
$0-$9,999			2.15 *	[1.17, 3.94]					2.14 *	[0.38, 1.90]
$10,000-$14,999			3.23 **	[1.55, 6.71]					3.31 **	[1.57, 6.97]
$15,000+			3.10 **	[1.43, 6.70]					3.26 **	[1.50, 7.09]
Difference in % of vacant housing units
<0.00%					1.17	[0.61, 2.25]			1.23	[0.63, 2.39]
0.00–2.99%					1.45	[0.78, 2.70]			1.39	[0.73, 2.62]
3.00–4.99%					1.86	[0.85, 4.04]			1.87	[0.86, 4.06]
5.00%+					1.00				1.00
Difference in median housing value
<$100,000							1.00		1.00
$100,000-$149,999							1.12	[0.60, 2.10]	0.92	[0.49, 1.70]
$150,000-$249,999							0.94	[0.51, 1.74]	0.76	[0.41, 1.40]
$250,000+							1.32	[0.55, 3.20]	0.98	[0.39, 2.43]

Note. OR = odds ratio; CI = confidence interval. The boldfaced values indicate statistical significance at the level of p < .05.

^*p< .05.

^**p < .01.

^***p < .001.

Poor families in neighborhoods that experienced an increase in median household income by less than $10,000, $10,000 to 14,999, or $15,000 or more had higher odds of having fruits in the home very often compared with those in neighborhoods that experienced a decrease in median household income (ORs = 2.04, 3.32, and 2.88, respectively, see Model 2, Table 3). When including all the three variables of neighborhood economic change in the analysis, an increase in median household income by $10,000 to 14,999 and $15,000 or more remained significant (see Model 5). Difference in the proportion of vacant housing units and difference in median housing value were not significantly associated with availability of fruits in the home among poor families.

Table 3.

Logistic Regression Analysis Assessing Associations Between Changes in Neighborhood Economic Indicators After the Great Recession and Availability of Fruits in the Home Among Poor Families, Geographic Research on Wellbeing (GROW) Study, California, United States, 2012–2013, n = 575.

	Model 1		Model 2		Model 3		Model 4		Model 5
Characteristic	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI
Mother’s age	1.03	[1.00, 1.07]	1.04 *	[1.00, 1.08]	1.03	[0.99, 1.07]	1.03	[0.99, 1.07]	1.04	[0.99, 1.08]
Mother’s race/ethnicity
Black	0.47	[0.13, 1.72]	0.51	[0.13, 1.95]	0.46	[0.13, 1.70]	0.50	[0.14, 1.84]	0.51	[0.13, 1.99]
Latina, foreign-born	0.34	[0.11, 1.02]	0.32 *	[0.10, 0.98]	0.33 *	[0.11, 0.99]	0.34	[0.11, 1.05]	0.32	[0.10, 1.01]
Latina, U.S.-born	0.45	[0.13, 1.54]	0.41	[0.11, 1.46]	0.44	[0.13, 1.54]	0.44	[0.13, 1.52]	0.40	[0.11, 1.46]
White, non-Hispanic	1.00		1.00		1.00		1.00		1.00
Mother’s marital status
Married or living together	1.38	[0.76, 2.51]	1.18	[0.65, 2.17]	1.31	[0.71, 2.41]	1.34	[0.74, 2.44]	1.13	[0.61, 2.10]
Unmarried	1.00		1.00		1.00		1.00		1.00
Number of children in household
1	1.00		1.00		1.00		1.00		1.00
2	1.44	[0.59, 3.51]	1.32	[0.52, 3.34]	1.49	[0.60, 3.69]	1.54	[0.62, 3.82]	1.41	[0.53, 3.70]
3+	2.07	[0.88, 4.88]	2.07	[0.84, 5.09]	2.17	[0.89, 5.27]	2.14	[0.89, 5.11]	2.14	[0.82, 5.57]
Mother’s education
Less than high school	1.00		1.00		1.00		1.00		1.00
High-school graduate/GED	1.70 *	[1.02, 2.81]	1.66 *	[1.00, 2.75]	1.67	[1.00, 2.78]	1.67 *	[1.01, 2.78]	1.62	[0.97, 2.71]
College or above	1.07	[0.40, 2.86]	1.00	[0.37, 2.73]	1.07	[0.39, 2.97]	1.02	[0.37, 2.79]	0.99	[0.34, 2.88]
Mother’s employment status
Did not work	1.00		1.00		1.00		1.00		1.00
Part-time work	1.30	[0.56, 3.02]	1.33	[0.56, 3.17]	1.23	[0.53, 2.87]	1.27	[0.54, 2.99]	1.25	[0.51, 3.05]
Full-time work	1.12	[0.66, 1.88]	1.12	[0.66, 1.91]	1.09	[0.64, 1.85]	1.07	[0.64, 1.81]	1.07	[0.62, 1.83]
Difference in median household income
<$0			1.00						1.00
$0-$9,999			2.04 *	[1.07, 3.90]					1.92	[0.99, 3.73]
$10,000-$14,999			3.32 **	[1.52, 7.27]					3.13 **	[1.42, 6.94]
$15,000+			2.88 **	[1.30, 6.39]					2.70 *	[1.19, 6.12]
Difference in % of vacant housing units
<0.00%					0.96	[0.47, 1.93]			0.99	[0.48, 2.03]
0.00–2.99%					1.22	[0.62, 2.39]			1.13	[0.55, 2.31]
3.00–4.99%					1.69	[0.73, 3.93]			1.73	[0.74, 4.06]
5.00%+					1.00				1.00
Difference in median housing value
<$100,000							1.00		1.00
$100,000-$149,999							1.57	[0.83, 2.99]	1.33	[0.69, 2.56]
$150,000-$249,999							1.17	[0.63, 2.18]	1.00	[0.53, 1.91]
$250,000+							1.97	[0.81, 4.79]	1.54	[0.61, 3.88]

Note. OR = odds ratio; CI = confidence interval. The boldfaced values indicate statistical significance at the level of p < .05.

^*p < .05.

^**p < .01.

^***p < .001.

As shown in Model 4, Table 4, living in neighborhoods that experienced an increase in median housing values by $150,000 to 249,999 had lower odds of cooking dinner at home frequently compared with the reference group among poor families (OR = 0.47). An increase in median housing values by $150,000 to 249,999 remained marginally significant at α = .10 when including all the three variables of neighborhood economic change in the analysis (see Model 5). Differences in median household income and proportion of vacant housing units were not significantly associated with cooking dinner at home among poor families. There were no significant associations between the changes in neighborhood economic indicators and home food environments among nonpoor families (see Supplemental Appendix Tables A1–A3).

Table 4.

Logistic Regression Analysis Assessing Associations Between Changes in Neighborhood Economic Indicators After the Great Recession and Cooking Dinner at Home for 5–7 Days a Week Among Poor Families, Geographic Research on Wellbeing (GROW) Study, California, United States, 2012–2013, n = 575.

	Model 1		Model 2		Model 3		Model 4		Model 5
Characteristic	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI
Mother’s age	0.97	[0.93, 1.01]	0.97	[0.93, 1.01]	0.97	[0.93, 1.01]	0.97	[0.94, 1.01]	0.97	[0.94, 1.01]
Mother’s race/ethnicity
Black	0.71	[0.30, 1.66]	0.67	[0.28, 1.57]	0.69	[0.29, 1.65]	0.70	[0.29, 1.69]	0.65	[0.26, 1.63]
Latina, foreign-born	2.74 **	[1.31, 5.71]	2.77 **	[1.32, 5.81]	2.62 *	[1.24, 5.56]	2.91 **	[1.37, 6.20]	2.77 **	[1.29, 5.99]
Latina, U.S.-born	1.29	[0.53, 3.14]	1.30	[0.53, 3.18]	1.29	[0.53, 3.17]	1.32	[0.53, 3.28]	1.33	[0.53, 3.34]
White, non-Hispanic	1.00		1.00		1.00		1.00		1.00
Mother’s marital status
Married or living together	1.78 *	[1.02, 3.10]	1.78 *	[1.03, 3.07]	1.80 *	[1.03, 3.15]	1.86 *	[1.06, 3.27]	1.84 *	[1.06, 3.20]
Unmarried	1.00		1.00		1.00		1.00		1.00
Number of children in household
1	1.00		1.00		1.00		1.00		1.00
2	0.71	[0.31, 1.59]	0.74	[0.32, 1.67]	0.73	[0.32, 1.66]	0.72	[0.32, 1.63]	0.78	[0.34, 1.81]
3+	1.48	[0.68, 3.23]	1.55	[0.70, 3.44]	1.52	[0.69, 3.35]	1.43	[0.65, 3.14]	1.55	[0.69, 3.53]
Mother’s education
Less than high school	1.00		1.00		1.00		1.00		1.00
High-school graduate/GED	0.99	[0.54, 1.79]	1.00	[0.55, 1.82]	0.97	[0.53, 1.77]	1.02	[0.56, 1.85]	0.99	[0.54, 1.80]
College or above	1.32	[0.53, 3.28]	1.38	[0.56, 3.37]	1.33	[0.53, 3.35]	1.47	[0.58, 3.69]	1.51	[0.60, 3.80]
Mother’s employment status
Did not work	1.00		1.00		1.00		1.00		1.00
Part-time work	0.64	[0.27, 1.52]	0.70	[0.28, 1.73]	0.62	0.26–1.46	0.68	0.28–1.63	0.69	0.28–1.72
Full-time work	0.36 ***	[0.21, 0.61]	0.36 ***	[0.21, 0.61]	0.36 ***	[0.21, 0.61]	0.37 ***	[0.22, 0.63]	0.36 ***	[0.21, 0.62]
Difference in median household income
<$0			1.00						1.00
$0-$9,999			1.19	[0.55, 2.54]					1.24	[0.57, 2.73]
$10,000-$14,999			1.22	[0.51, 2.89]					1.31	[0.53, 3.21]
$15,000+			0.62	[0.27, 1.43]					0.74	[0.31, 1.75]
Difference in % of vacant housing units
<0.00%					1.47	[0.70, 3.08]			1.56	[0.74, 3.27]
0.00–2.99%					1.24	[0.63, 2.44]			1.42	[0.71, 2.84]
3.00–4.99%					1.71	[0.74, 3.94]			1.89	[0.81, 4.43]
5.00%+					1.00				1.00
Difference in median housing value
<$100,000							1.00		1.00
$100,000-$149,999							0.65	[0.32, 1.34]	0.62	[0.29, 1.33]
$150,000-$249,999							0.47 *	[0.23, 0.96]	0.49	[0.23, 1.06]
$250,000+							0.54	[0.22, 1.32]	0.56	[0.21, 1.47]

Note. OR = odds ratio; CI = confidence interval. The boldfaced values indicate statistical significance at the level of p < .05.

^*p< .05.

^**p< .01.

^***p < .001.

Discussion

This study investigated neighborhood-level mechanisms by which the Great Recession affected home food environments. The findings partly support the idea that poor families residing in neighborhoods severely affected by the Great Recession were vulnerable to unhealthy home food environments after the Great Recession. The findings highlight the importance of the intersection between national, neighborhood, and family economic conditions in understanding home food environments. To our knowledge, this is the first study to link neighborhood economic changes after the Great Recession to home food environments with consideration of family-level poverty status.

Differential Neighborhood Economic Changes After the Great Recession

Neighborhood economic changes after the Great Recession were shown to be concentrated in socioeconomically disadvantaged neighborhoods (Allen, 2013; Delmelle & Thill, 2014; Hyra & Rugh, 2016; A. Owens & Sampson, 2013; Solari, 2012; Williams et al., 2013). This study showed a similar pattern in that neighborhoods with a high concentration of poverty had smaller increases in household median income and median housing values and a larger increase in proportions of vacant housing units between 2000 and 2009–2013 than those with a low concentration of poverty (see Figure 1). The vulnerability of economically disadvantaged neighborhoods to the recession is probably because many of their residents have unstable jobs sensitive to a macroeconomic crisis and had a high percentage of home mortgages (particularly subprime loans) in the early- to mid-2000s (Hyra & Rugh, 2016).

Neighborhood Economic Changes and Home Food Environments

Home food environments became increasingly unhealthy after the Great Recession (Anderson et al., 2014; Bruening et al., 2012; Griffith et al., 2013; Hall & Perry, 2013; U.K. Department for Environment, Food and Rural Affairs, 2012). A housing market collapse and increasingly unstable employment status brought about severe family economic hardship (Allen, 2013; Ellen & Dastrup, 2012; Grusky, Western, & Wimer, 2011; Palmer, 2013; Williams et al., 2013), which likely increased the risk of unhealthy home food environments (Adams et al., 2015; MacFarlane, Crawford, Ball, Savige, & Worsley, 2007). In addition to a family-level mechanism, the role of macroeconomic and neighborhood economic contexts in home food environments has been conceptually established (Rosenkranz & Dzewaltowski, 2008; Weiland & Yoshikawa, 2012), yet little empirical literature has investigated home food environments in macroeconomic and neighborhood economic contexts.

This study focused on differences in neighborhood economic context between before and after the Great Recession as a correlate of home food environments and showed that poor families living in neighborhoods with decreased median household income after the Great Recession had less availability of fruits and vegetables in the home compared with those in neighborhoods with increased median household income. A possible mechanism is that worsening neighborhood economic conditions after the Great Recession reduced residents’ ability to buy fresh produce (Letamendi, 2014), local government tax revenue which could be used for health-promoting services and facilities (Chernick et al., 2011; Weiland & Yoshikawa, 2012), and neighborhood social cohesion (L. A. Owens & Cook, 2013; Sampson et al., 1997), all of which are associated with unhealthy home food environments (Rosenkranz & Dzewaltowski, 2008; Weiland & Yoshikawa, 2012). Empirical findings on the associations among neighborhood economic, service, and social environments (Kim & Cubbin, 2017; Sallis et al., 2011) and the association between neighborhood social and service environments and home food environments or home food insecurity (Dean & Sharkey, 2011; Ding et al., 2012; D. B. Jackson et al., 2019; Nebbitt et al., 2016; Shier et al., 2016) have corroborated the mechanism. On the other hand, increases in neighborhood-level median housing values after the Great Recession were associated with lower odds of cooking dinner at home frequently among poor families. Neighborhoods in which housing values substantially increased might be gentrified ones which replaced older and local restaurants with newer restaurants and stores. Recent changes in neighborhood retail environments may explain an “eating-out” trend among poor families.

The situation was different for nonpoor families: Neighborhood economic change after the Great Recession was associated with neither availability of fruits and vegetables in the home nor frequent cooking of dinner at home. Nonpoor families may be able to avoid the harmful effects of neighborhood economic deterioration on home food environments by utilizing their family-level financial resources and social support. For example, they could afford healthy local and organic food at higher end supermarket chains outside their neighborhood, especially when the food retail market in their neighborhood may have fewer high-quality and expensive food ingredients. Also, they would be less affected by budget cuts in public health and nutrition services induced by the Great Recession compared with poor/near-poor families. Regarding cooking dinner at home frequently, nonpoor families might eat in restaurants and takeaways inside and outside of their neighborhood (regardless of their neighborhood’s retail environments) because they can afford it. Prior studies have also found nonsignificant associations between neighborhood contexts and health behaviors among nonpoor people (Kim & Cubbin, 2017; Lee, Kim, & Cubbin, 2018).

Limitations

This study has several limitations. The first is a cross-sectional design and not measuring change in home food environments between before and after the recession. Second, we could not examine the mechanisms through which change in neighborhood economic context influences home food environments. Conceptual frameworks suggest potential mediators in such associations—through reduced ability for consumption, government tax revenue, and neighborhood retail, service, and social environments (Chernick et al., 2011; Letamendi, 2014; L. A. Owens & Cook, 2013; Sampson et al., 1997; Weiland & Yoshikawa, 2012). Future research should explain the neighborhood-level mechanism of the recession not only to identify policy target variables but also to elaborate the reasons why neighborhood employment-related indicators in comparison to housing-related indicators were differently associated with home food environments. Third, mother-reported measures of home food environments could be biased by social desirability and recall bias. Fourth, although we controlled for some of family-level demographic and socioeconomic characteristics in the analysis, there are still factors that we did not account for in the study, such as public assistance, parental history of incarceration, neighborhood social cohesion, neighborhood safety, and community resources. Further research needs to include a more comprehensive set of covariates to investigate the neighborhood-level mechanisms by which the Great Recession affected home food environments. Fifth, our findings are limited to one state and urban areas, which cannot be generalized across the United States. Finally, census tracts, which we used to define neighborhoods can be different from resident-defined neighborhoods (Coulton, Korbin, Chan, & Su, 2001) although designed to encompass a homogeneous and distinctive geographic area.

Conclusion

Despite these limitations, this study has several strengths. This study confirmed perpetuated neighborhood-level economic disparities by finding that economic hardships induced by the Great Recession were distributed unequally. Perpetuated neighborhood economic disparities may negatively impact residents’ health in disadvantaged neighborhoods gradually over the next decades considering neighborhood impacts on individual health outcomes can take time to manifest (A. Owens & Sampson, 2013). Furthermore, food environments in a child’s home have been shown to influence the consumption of obesity-promoting foods and lower energy density alternatives such as fruits and vegetables (Arcan et al., 2007; Blanchette & Brug, 2005; Ding et al., 2012; Golan & Weizman, 2001; Hanson et al., 2005; J. A. Jackson et al., 2017; Larson et al., 2006; Neumark-Sztainer et al., 2003; Wyse et al., 2011) and other behavioral, emotional, cognitive, and academic development of children (Faught et al., 2017; D. B. Jackson, 2016; D. B. Jackson et al., 2018; D. B. Jackson & Vaughn, 2017; Johnson & Markowitz, 2018; Shankar et al., 2017). Understanding the influences on child holistic development, future research should investigate long-term trends in neighborhood economic disparities and their links to home food environment and subsequently various child developmental domains.

In addition, this study contributed to the public health literature by examining the intersection of national, neighborhood, and family economic status in understanding home food environments. The findings imply that home food environments are affected by interactions among macro- and mezzo-level economic contexts as well as individual-level socioeconomic status, cultural characteristics, and personal preference. Because national and neighborhood economic contexts serve as extra barriers to the adoption of micro-level interventions, public health interventions aiming to improve home food environments and promote children’s healthy development should include strategies at the national and neighborhood levels as well as the family level. This study also suggests that a main target population of public health interventions designed to promote children’s daily consumption of fruits and vegetables should be poor families living in disadvantaged neighborhoods. As an example of neighborhood-level strategy for poor families, improving accessibility to healthy food outlets and affordable healthy food options in poor neighborhoods can facilitate their residents’ purchase of healthful food ingredients (Jetter & Cassady, 2006; Story, Kaphingst, Robinson-O’Brien, & Glanz, 2008). In addition, efforts to improve accessibility to and affordability of healthful foods may be more effective when combined with nutrition assistance programs that provide nutrition education and food vouchers to poor families’ and their children’s diet. The family- and neighborhood-level interventions could make economically disadvantaged neighborhoods and their poor residents and their children more resilient in the face of a national economic crisis and also reduce socioeconomic disparity in child nutrition and health (Ranjit et al., 2015). Thus, public health, pediatric, behavioral, and developmental professionals should ensure that poor families in disadvantaged neighborhoods are connected with available public nutrition assistance programs which promote healthy food environments in households with children given the impact of home food environments on physical, behavioral, emotional, cognitive and academic development of children.

Finally, in contrast to availability of vegetables and fruits, poor families in neighborhoods in which the Great Recession had the least negative effect on their real-estate market had a low frequency of cooking food at home, suggesting that they eat out more often, and meals from restaurants and fast food outlets tend to be more calorie-dense and contain more saturated fat and less dietary fiber, calcium, and iron than at-home food (Guthrie, Lin, & Frazao, 2002). Caloric and fat nutrition information printed on menus and relative caloric information in the form of physical activity equivalents would increase customer awareness and possibly prevent overconsumption outside the home (Bleich, Herring, Flagg, & Gary-Webb, 2012).