Neighborhoods and Food Insecurity in Households with Young Children: A Disadvantage Paradox?

Neighborhoods and Food Insecurity in Households with Young Children: A Disadvantage Paradox? Abstract In the United States, more than 1 in 5 households with children are unable to access and provide adequate food for a healthy, active lifestyle. We argue that the contribution of local context for food insecurity risk has largely been overlooked in favor of focusing on individual family characteristics, and that this is problematic given that mitigating food insecurity may be a communal process. We examine the relevance of neighborhood contributors to food insecurity among children, utilizing geocoded and nationally-representative data from the ECLS-K: 2010-2011 kindergarten cohort. We find little evidence that neighborhood socioeconomic, food retail, or social services characteristics directly impact food insecurity risk. However, our results reveal that family and neighborhood socioeconomic characteristics combine to impact food insecurity in ways consistent with a disadvantage paradox. As neighborhood concentrated disadvantage increases, higher-SES families’ risk of food insecurity increases, but lower-SES families’ risk decreases. This paradox is not explained by a higher concentration of social service organizations in more disadvantaged neighborhoods, and we theorize that impoverished families with children may share information and resources in disadvantaged communities to avoid food insecurity. food insecurity, SES, neighborhood disadvantage, ECLS-K, children Researchers, policymakers, and citizens alike agree that all children deserve the opportunity to develop and thrive. However, in the United States, one of the wealthiest nations in the world, 1 in 5 children live in poverty, including 24% of all children under the age of 6 (CLASP 2013). Poverty during childhood has immediate and long-term consequences for educational achievement, employment and earnings prospects, and health. Many children growing up in poverty experience these deficits as a result of living in a household that struggles with having enough food. Household food insecurity, a household’s collective inability to access adequate food for a healthy, active lifestyle, impacts nearly 16% of all households in the U. S. (Coleman-Jensen, Gregory and Singh 2014) and more than one in five households with children (Wight, Thampi and Briggs 2010). More troubling in this post-economic recession era, food insecurity is at the highest level of severity ever measured and has increased over 30% since 2007 despite federal food and nutrition assistance programs aimed at its elimination (Nord et al. 2010). Moreover, it is likely that estimates of the prevalence of food insecurity among households with children are underestimates – given that parents may be reluctant to admit their inability to provide enough food for their families, or are unaware that their children are suffering. One study of adolescents showed wide discrepancies between parent- and adolescent-reported accounts of food insecurity (Nord and Hanson 2012), and another mixed-methods study found a similar pattern for children’s accounts of food insecurity (Fram et al. 2013), such that adolescents and children report being food insecure at higher rates than do their parents. The precise nature of how to identify, examine, and alleviate food insecurity is complex. Strategies and current policies to address food insecurity among children have largely focused on families. While these policies may help individual families, to date they have struggled to curtail this enduring problem. We argue that one reason behind stalled progress may be failing to consider the community characteristics that might influence food insecurity, above and beyond individual or family factors. Existing knowledge of who and especially where food insecure children are remains quite limited, as well as how family and community characteristics might combine to influence the risk of food insecurity. We have three primary aims in this paper. First, we document how the neighborhoods of food secure and insecure children differ. Second, we investigate whether and how neighborhood characteristics influence the odds of food insecurity for these households by simultaneously considering the influence of family and neighborhood level characteristics. Finally, we assess how neighborhood characteristics might moderate the powerful influence of family-level SES on a child’s risk of living in a food insecure household. Food Insecurity Among Children Scholars interested in poverty and its effects on children should understand that household food insecurity is an indicator of material hardship that has far-reaching implications for children’s development and is social in nature. That is, it is a factor facing families which may be alleviated by collective and reciprocal assistance from entities outside the home: extended family members, other families, neighbors, or social and governmental organizations. Perhaps due in part to collective processes, poverty and food insecurity are strongly related but they are not synonymous. This is evidenced by the many poor families who remain food secure (Gundersen, Kreider, and Pepper 2011). Thus, for scholars interested in social processes related to disadvantage and how it might be mitigated, food insecurity is critical to understand. In this paper, we focus on food insecurity among households with children. Typically, the percentage of households with children who report food insecurity exceeds the percentage of food insecure children because in some households, parents will forego meals to ensure their children receive adequate food. Moreover, qualitative evidence demonstrates that even when parents believe they are shielding their children from the detrimental effects of food insecurity, children still suffer and in fact, take responsibility for implementing complex strategies of their own to mitigate food insecurity for their families (Fram et al. 2011). The consequences of food insecurity for children’s well-being are clearly established (Alaimo et al. 2001; Cook et al. 2004; Gundersen, Kreider and Pepper 2011). Existing research shows that the number of episodes of hunger that children experience is related to their health as they grow (Kirkpatrick, McIntyre and Potestio 2010), indicating that the healthy development of children is associated with not just the presence, but also the severity of food insecurity. Children who experience persistent food insecurity during early childhood, for instance, have worse health in adolescence relative to those who never or transiently experience food insecurity (Ryu and Bartfeld 2012). The costs of food insecurity in children extend beyond physical to mental health and academic performance. Detrimental performance in math and reading, loss of school days and repeated grades, behavior or attention problems, special education or mental health counseling, and suicidal or depressive tendencies among adolescents have all been linked to living in food insecure households (Alaimo, Olson and Frongillo 2002; Alaimo, Olson and Frongillo 2001; Jyoti, Frongillo and Jones 2005; Kimbro and Denney 2015; Kleinman et al. 1998; McIntyre et al. 2013; Murphy et al. 1998; Whitaker, Phillips and Orzol 2006). Much research has focused on the household-level characteristics associated with food insecurity. Family structure is an important predictor, with increased risk among households headed by single women (Nord et al. 2010). Maternal factors, including mental health and citizenship statuses (Van Hook and Balistreri 2006), also affect children’s risk of food insecurity. For example, children who have noncitizen mothers have levels of food insecurity nearly twice as high as those with native-born mothers (Kalil and Chen 2008), likely partially due to the families’ ineligibility (or belief of ineligibility) for federal food assistance. In addition, rates of food insecurity are higher than the national average among black and Hispanic households (Nord et al. 2010). As with a number of adult and child well-being outcomes, low socioeconomic status (SES), as indicated through a variety of measures, such as maternal education, family income, and participation in food assistance programs such as the Supplemental Nutrition Assistance Program (SNAP; formerly known as Food Stamps), WIC, or free/reduced breakfast and lunch, are strongly associated with household food insecurity, even after accounting for issues of selection (Alaimo et al. 1998; Casey et al. 2001; Jones et al. 2003; Kalil and Chen 2008; Rose and Richards 2004). However, at the same time, researchers have puzzled over how some disadvantaged families remain food secure (Nord et al. 2010). What strategies are poor families who remain food secure implementing? How are they shielding their children from harm? Below, we invoke interdisciplinary scholarship on poverty and discuss how some poor families may be relying on other external sources of resource allocation and support to avoid food insecurity, and argue that access to these external sources of assistance is not evenly distributed across communities. Neighborhoods and Well-being Though characteristics of children and their families motivate most research and policy strategies to curb food insecurity to date, there is ample evidence that neighborhood factors influence child well-being (Leventhal and Brooks-Gunn 2000). We conceptualize communities as influencing the prospects for healthy development because of the institutional and social resources immediately available (or unavailable) to residents (Jencks and Mayer 1990). Neighborhoods with more wealth tend to attract more institutional resources such as retail establishments, parks, and other amenities. Higher-SES neighborhoods also may be characterized by a greater density of other institutional resources like schools, libraries, and community organizations, the lack of which may contribute to the impact of neighborhood disadvantage on residents’ well-being (Wilson 1996). The social resources available in a neighborhood, or community social organization, may also impact food insecurity in a variety of ways (Crowe and Smith 2012). On the one hand, more advantaged areas may have higher levels of social integration, provide more secure and beneficial social networks, an enhanced sense of social capital for residents, and consequently represent places where children are more likely to thrive (Sampson 2012). Neighborhoods rich in social ties and with norms of reciprocity are likely to be places where neighbors look out for each other. In contrast, neighborhoods where distrust pervades the community are places where social capital has difficulty being generated or sustained, and neighbors may keep to themselves (Rankin and Quane 2000). Social resources can take at least two forms in communities – ties to other individuals or families, and ties to social organizations. In more disadvantaged neighborhoods, residents are less likely to have employed friends and more likely to have friends on public assistance (Rankin and Quane 2000), and better-off residents may minimize their associations with lower-income neighbors (Anderson 1990). Ties to civic institutions and social organizations may also be diminished in some (Wilson 1987), though importantly not all (Klinenberg 2002; Small 2004), lower-income neighborhoods. In all, social and economic disadvantages may extend beyond individual or family disadvantages to area-level disadvantages to influence well-being. On the other hand, linking neighborhood socioeconomic disadvantage to individual well-being is more complicated. Though there are several examples to suggest that poor residents in poor neighborhoods face special heightened threats to well-being (Wilson 1987; Wilson 1996), scholars have shown time and again that not all poor communities are alike (Klinenberg 2002; Seccombe 2002; Sherman 2006; Small 2004; Small, Harding and Lamont 2010). In spite of depictions of poor communities as devoid of social capital with residents destined for poor health and well-being, there are clear exceptions that reveal strategies to persevere in the face of seemingly overwhelming adversity (Desmond 2012; Klinenberg 2002; Seccombe 2002; Small 2004; Stack 1974). Accordingly, there is value in assessing whether well-being differs for more and less disadvantaged families who reside in neighborhoods of different levels of socioeconomic status. Neighborhoods and Food Insecurity To date, social scientists have paid little attention to the relationship between social and economic neighborhood environments, including the availability of food and social service agencies, and food insecurity. As a notable exception, Kirkpatrick and Tarasuk’s (2010; 2011) research evaluates the importance of both food environments and perceived neighborhood social capital on families’ food insecurity risks in an urban area in Canada. The authors assess the distance to food outlets and neighborhood social capital and conclude that the causes of food insecurity lie with characteristics of households and not necessarily neighborhoods. These results stand in contrast to neighborhood effects literatures (see Kawachi and Berkman 2003) which suggest that area-level characteristics shape individual well-being. A recent descriptive U.S. study suggests that structural factors at the neighborhood level, such as aggregate levels of poverty, may impact food insecurity above and beyond household-level factors (Kimbro, Denney and Panchang 2012). Neighborhood disadvantage as it may relate to household food insecurity can be described in multiple ways. First, social and economic characteristics such as the neighborhood proportion in poverty, proportion unemployed, proportion of female-headed households, and the proportion receiving public assistance collectively represent variation in what has been termed concentrated disadvantage across neighborhoods (Morenoff 2003; Sampson, Raudenbush and Earls 1997; Swaroop and Morenoff 2006). Second, food availability in neighborhoods can be measured more directly by assessing the neighborhood availability of supermarkets and convenience stores. One assumption is that a greater availability of grocery stores could reduce food insecurity, particularly in low-income neighborhoods with few or no supermarkets. Conversely, a high representation of convenience stores might indicate a lack of quality foods. The literature on the link between grocery store access and food security is international in scope and mixed in findings, with some studies reporting positive associations and others reporting no association (see Sadler, Gilliland and Arku 2013 for a discussion). Recent work in Flint, Michigan finds no significant association between physical access to food and household food security (Sadler, Gilliland and Arku 2013). Similarly, another recent U.S. study in Philadelphia found no association between grocery store availability and obesity (Cummins, Flint and Matthews 2014). The recent evidence suggests that food access environments may not be as influential on populations as once perceived (Sadler, Gilliland and Arku 2013). However, these recent studies focused on single communities, so it remains unclear if food access nationwide is a significant predictor of household food insecurity. Finally, measurement of social services such as the availability of food pantries within neighborhoods may associate with household food insecurity over and above family level characteristics. One possibility is that social services aimed at alleviating food insecurity may be more prevalent in disadvantaged neighborhoods featuring residents most in need of assistance. A first set of hypotheses investigates the relationship between neighborhood indicators of disadvantage and household food insecurity over and above the relationships between family status and food insecurity: H1: Neighborhood concentrated disadvantage will be associated with higher odds of household food insecurity. H2: Greater supermarket availability in neighborhoods will be associated with lower, and higher proportions of convenience stores will be associated with higher, odds of household food insecurity. H3: Greater representation of social assistance organizations will be associated with lower odds of household food insecurity. These hypotheses assess if characteristics of the communities where children live influence the odds of food insecurity for the household but they do not specify how, and for whom, neighborhoods might matter. Below, we describe at least two possibilities aimed at understanding how neighborhoods might matter specifically for children in the most disadvantaged families: 1) accumulating household and neighborhood disadvantages and 2) a disadvantage paradox, whereby the most disadvantaged households may experience lower risks of food insecurity the more disadvantaged the neighborhood. Accumulating Risks Decades of neighborhood research on important social problems such as health (Kawachi and Berkman 2003), safety (Sampson, Raudenbush and Earls 1997), and educational attainment (Borman and Dowling 2010; Jencks and Mayer 1990), stipulate that disadvantages across individuals and the places they live possess the capability to accumulate. For example, a recent and sophisticated study demonstrated that living in a disadvantaged neighborhood is most consequential for high school graduation for poor children (Wodtke, Harding and Elwert 2016). In the context of food insecurity, an accumulation of risk might include places with few jobs, little transportation, and/or few supermarkets that provide healthy foods at lower costs and which accept assistance programs such as SNAP. These same places disproportionately include individuals and families that possess fewer resources who would rely more heavily on neighborhood resources were they available. Thus, the contributions of household and neighborhood level disadvantages might synergistically influence risk. If so, then neighborhood disadvantages increase the risk of food insecurity more for more socioeconomically disadvantaged households than for less disadvantaged households. Thus, H4: The risk of food insecurity for children in the most disadvantaged families will increase the more disadvantaged the neighborhood. A Disadvantage Paradox Conceptualizing food insecurity, especially for children, as an outcome with collective or community implications reveals the possibility of a unique interaction between neighborhoods and families. Not all poor communities are alike (Crowe and Smith 2012; Klinenberg 2002; Small, Harding and Lamont 2010). Structurally, some poor neighborhoods possess resources such as community centers or food pantries specifically aimed at alleviating, and in response to, disadvantage. These community resources provide the potential to help all residents, regardless of individual SES (Carpiano, Lloyd and Hertzman 2009), and some evidence suggests that participation in social organizations may be higher in more disadvantaged neighborhoods than in less disadvantaged ones (Rankin and Quane 2000). Further, neighborhood characteristics comprise indicators of social networks and connectivity that come with links to resources or information (Carpiano 2008; Kawachi 2010; Small 2009; Swaroop and Morenoff 2006). Given that healthy foods are not equally accessible across all communities (Hung 1999; Jetter and Cassady 2006; Morland, Diez Roux and Wing 2006), community resources and support that can be leveraged to avoid food insecurity may also fluctuate. Qualitative research concludes that insufficient food supplies are likely not the result of some deficiency at the household level but rather extend to wider social networks and the social and economic characteristics of the communities in which families live (Ahluwalia, Dodds and Baligh 1998). There is also evidence to suggest that greater civic structure within communities can reduce the chance of food insecurity (De Marco and Thorburn 2009; Morton et al. 2005; Vozoris and Tarasuk 2003). A small study in Connecticut found that social capital at the household and community level was associated with a reduced risk of food insecurity (Martin et al. 2004). Indeed, researchers have documented that disadvantaged persons and families pool resources with each other to endure difficult circumstances (Morton et al. 2008; Stack 1974). Individuals in poor communities will respond to deprivation in varying ways as well, influenced in part by cultural expectations and moral rules informally and formally enacted by area residents (Anderson 1999; Sherman 2006). Coping strategies in poor neighborhoods that are viewed as more in line with expectations of behavior may generate what Jennifer Sherman (2006) refers to as “moral capital.” In turn, moral capital can be traded for “…social capital in the form of community ties or social support” (Sherman 2006: 893). Though Sherman identifies this form of social resource sharing in rural communities dealing with poverty, Lichter and Brown (2011) provide reasons to believe that processes occurring in rural spaces might under some circumstances also occur in urban settings as the grip of poverty does not discriminate spatially, nor do collective reactions to it. Recent research on the working poor and the Earned Income Tax Credit (EITC) bears this out. Indeed, poor residents almost uniformly reject the idea that the EITC is a form of government assistance but rather, as the name implies, earned income (Halpern-Meekin et al. 2015). Thus, applying for and receiving the EITC to help with ongoing economic struggles is not only a survival strategy but perhaps the moral and right thing to do, creating a cultural space where bonds are formed and values shared. In addition, disadvantaged groups at times exhibit more prosocial resource allocation and a greater tendency toward egalitarian social values than more advantaged persons and families (Piff et al. 2010). In the case of households with young children, these phenomena may be particularly pronounced. Establishing strong networks can be seen as an adaptive strategy used to deal with the consequences of poverty (Dominguez and Watkins 2003; Menjivar 2000; Newman 1992). Although many scholars believe the kin networks Stack (1974) observed have all but disappeared in the modern disadvantaged neighborhood (see Smith 2007), Desmond (2012) writes that they have been replaced with quickly-formed and quickly-dissolved intense ties which enable survival. Finally, evidence suggests that if these processes occur in disadvantaged neighborhoods they are likely to apply most to the most disadvantaged residents of the area. Anderson (1990) witnessed patterns of behavior in disadvantaged neighborhoods where more advantaged residents minimized their associations with lower-income neighbors. It may be that better-off residents do not agree with the coping strategies employed by their struggling counterparts and thus remove themselves from interactions that might generate moral and social capital (Sherman 2006). This stream of research provides an alternative to the accumulation of disadvantage hypothesis: H5: The risk of food insecurity for children in the most disadvantaged families will decrease the more disadvantaged the neighborhood. By comparison, for both the accumulation of risks and the disadvantage paradox perspectives, children in less disadvantaged families may experience higher risks of food insecurity the more disadvantaged the neighborhood. But if the accumulation perspective is supported, then those increases will be less pronounced for children in less disadvantaged families compared to children in the most disadvantaged families. Alternatively, if the paradox is supported, then we should see a reduction in risk as neighborhood disadvantage increases for more disadvantaged children; and an increase in risk for less disadvantaged children. DATA AND METHOD This study uses restricted, geo-coded data from the spring kindergarten wave of the Early Childhood Longitudinal Study-Kindergarten Class of 2010-2011 (ECLS-K), which is a nationally-representative sample of U.S. children who were in kindergarten in 2010-2011. Food insecurity was collected only in the spring data collection, so we utilize data primarily from the spring kindergarten wave. The restricted version of ECLS-K provides census tract information with which data from the American Community Survey (2007-2011) are appended to create neighborhood-level measures. We utilize the ACS instead of decennial Census 2010 now that the ACS routinely collects long-form information on socioeconomic characteristics such as poverty status. We also link the ECLS-K to 2010 Zip Code Business Patterns data to create estimates of the density of food retail and social service organizations in census tracts.1 The ECLS-K sample includes 18,100 children (in accordance with our restricted data agreement we round all sample sizes to the nearest 50) with valid geocodes at the spring kindergarten wave. About 29% of this sample is missing a parent interview at the spring wave, when food insecurity was collected, so our final sample size is 12,800 children who live in 3,800 Census tracts. Roughly 10% of the sample represents the sole observation in their census tract and the average number of children per census tract is 3.4; we find no evidence this creates estimation problems (Bell, Ferron and Kromrey 2008). But we do conduct sensitivity analyses by comparing our results before and after dropping the singleton neighborhood cases and find no differences in results so they are included in the final models. Variables Our outcome measure is a dichotomous measure of household food insecurity derived from the USDA’s 18-item food insecurity scale (Bickel et al. 2000). Examples of the questions include, “In the last 12 months, were you ever hungry but didn’t eat because you couldn’t afford enough food?” and “In the last 12 months, was [child] ever hungry but you just couldn’t afford more food?” Following USDA guidelines (2000), we classify the households of parents answering in the affirmative to three or more of the items as food insecure. Individual-level variables include the child’s age in months, race/ethnicity (non-Hispanic white, non-Hispanic black, non-Hispanic other, and Hispanic), and gender (1 = male); the mother’s age and nativity status (1 = foreign born); the number of siblings in the household and family structure (two-parent family, single-mother family, and other family type). We incorporate a series of health and well-being measures, which are asked of the primary caregiver of the child (the biological mother in most cases). The ECLS-K utilizes an abbreviated version of the 20-item Center for Epidemiological Studies of Depression Scale (CES-D) (Radloff 1977). Following convention, we classify mothers as “likely depressed” if her total score on the abbreviated scale is equal to or greater than 12. We include dichotomous indicators for whether the child and primary caregiver are in “fair or poor health,” utilizing the standard five-response self-rated health measure, as well as an indicator for whether the primary caregiver reports health-related work limitations. We also account for whether the household currently receives SNAP benefits, as well as the length of time the child has lived in the current residence (measured in years) – note that this measure was collected only in the fall kindergarten wave. To measure family SES, we utilize the composite measure provided by the ECLS, which incorporates household income, both parents’ occupational prestige, and both parents’ education (this measure incorporates the custodial parent’s data if only one parent is present in the household). Each component is standardized, and the resulting z-scores are added and divided by the number of components to yield an overall standardized measure of SES. To mirror interpretation for our neighborhood disadvantage measure described below, we reverse-coded this measure so that higher values indicate greater socioeconomic adversity. In other words, a value of 1 indicates a family is one standard deviation higher in socioeconomic adversity than the mean. To capture the neighborhood characteristics that may influence food insecurity risk we control for neighborhood population (logged) and a dichotomous indicator for whether the child lives in an urban area. For social and economic disadvantage in neighborhoods we include an index of concentrated disadvantage, created based on the first dimension of a principal components factor analysis on percent of adults in the census tract living below the poverty line, the percent of households receiving public assistance, the percent of adult residents who are unemployed, and the percent of female-headed households with children (Sampson, Raudenbush and Earls 1997). We also include two measures of the food retail environment in a neighborhood, the density of large supermarkets (50+ employees) and convenience stores per square mile. Finally, we include a measure of the density of social service organizations in the neighborhood, all measured per square mile – social assistance organizations related to children and youth (e.g. Boys and Girls Clubs, AFDC, day care), community services (e.g. housing shelters, food pantries), and vocational rehabilitation services (e.g., job training and counseling). These measures are also logged in our models. Missing Data Approximately 20% of children remaining in our analytic sample are missing data on one or more measures of interest, with the vast majority (80%) of those missing only on the number of years the child had lived at the current residence (because it was only collected in the fall kindergarten wave, so children missing a parent interview in the fall are missing on this item). Children missing data were more often non-Hispanic black or Hispanic, poorer, lived with single mothers, and were more likely to have foreign-born parents. Given the evidence that our missing data are not missing completely at random and may be conditioned by other observed covariates, standard procedures such as listwise deletion would be inappropriate (Allison 2002). Instead, we use multiple imputation procedures in Stata 12 (Royston 2005) to estimate values for our multivariate analyses. During imputation, a diverse set of predictors estimate five sets of probable values for each missing value. The resulting five data sets include a random component based on draws from the posterior predictive distribution of the missing data under a posited Bayesian model and provide unbiased estimates of variance (Allison 2002). Models estimated without imputation provide results very similar to those using imputation (available upon request). Estimation To test the effects of neighborhood conditions on odds of food insecurity among households with children, we estimate multi-level logistic regression models (Guo and Zhao 2000; Rabe-Hesketh and Skrondal 2008) using the mi estimate command within Stata 12 software (StataCorp 2012). The models treat level-1 children as nested within level-2 census tracts. All models utilize maximum likelihood estimation with adaptive quadrature (Rabe-Hesketh and Skrondal 2008), adjusting for clustering by neighborhood, different sample sizes for level-1 and level-2 units, heteroscedastic error terms, and varying numbers of cases within level-2 units – all problems that otherwise downwardly bias estimated standard errors (Raudenbush and Bryk 2002). We evaluate hypotheses 1 through 3 by including neighborhood-level (level-2) predictors one at a time along with the level-1 predictors. Hypotheses 4 and 5 specify cross-level interactions between family SES at level-1 and concentrated disadvantage, food retail, and social organizations at level-2. The error terms associated with the interaction models are assumed to be multivariate normally distributed, each with a mean of 0 and non-zero variances and covariances. We report all regression results as odds ratios (OR) and generate predicted probabilities to illustrate findings from the cross-level interactions. RESULTS First, Table 1 provides weighted means and proportions for the dependent and independent variables at both the individual/family level and at the neighborhood level. In addition, the table provides significance tests of differences for households reporting food insecurity versus those not reporting food insecurity. Table 1. Means and Proportions, ECLS-K: 2010-2011 Data Panel A. Individual and Family Characteristics n = 12,800 Full Sample Food Secure Food Insecurea Food insecure 0.12 Child is male 0.51 0.51 0.53 Child's age in months 74.60 74.50 74.80* Foreign-born parent 0.34 0.33 0.44*** Race/ethnicity  non-Hispanic white (ref) 0.51 0.54 0.34***  non-Hispanic black 0.11 0.10 0.15***  Hispanic 0.24 0.21 0.40***  Other 0.14 0.15 0.11*** Mother's age in years 34.50 34.70 33.20*** Family Structure  Two biological parents, married or cohab. (ref) 0.72 0.74 0.57***  Single mother family 0.20 0.18 0.33***  Other family type 0.08 0.08 0.10+ Family socioeconomic adversity (standardized) 0.05 −0.09 0.61*** Family receives food stamps 0.26 0.22 0.54*** Number of siblings in the household 1.46 1.42 1.75*** Length of time child has lived in current house (years) 2.77 2.81 2.49*** Child in fair/poor health 0.02 0.02 0.06*** Primary caregiver likely depressed 0.08 0.06 0.24*** Primary caregiver in fair/poor health 0.10 0.08 0.25*** Primary caregiver has health-related work limitations 0.07 0.06 0.16*** Panel B. Neighborhood Characteristics n = 3,800 Full Not Food Insecure Food Insecurea Neighborhood is urban 0.77 0.76 0.82*** Concentrated Disadvantage (standardized) 0.00 −0.13 0.33***  (components of concentrated disadvantage)   % Households in poverty 0.13 0.12 0.16***   % Female headed household 0.14 0.13 0.16***   % unemployed 0.09 0.08 0.10***   % Residents receiving public assistance 0.03 0.02 0.03*** Neighborhood Resource Measures (density per square mile)  Large supermarkets (50+ employees) 0.28 0.26 0.29*  Convenience Stores 0.46 0.39 0.53***  Social assistance organizations 2.89 2.47 3.13** Panel A. Individual and Family Characteristics n = 12,800 Full Sample Food Secure Food Insecurea Food insecure 0.12 Child is male 0.51 0.51 0.53 Child's age in months 74.60 74.50 74.80* Foreign-born parent 0.34 0.33 0.44*** Race/ethnicity  non-Hispanic white (ref) 0.51 0.54 0.34***  non-Hispanic black 0.11 0.10 0.15***  Hispanic 0.24 0.21 0.40***  Other 0.14 0.15 0.11*** Mother's age in years 34.50 34.70 33.20*** Family Structure  Two biological parents, married or cohab. (ref) 0.72 0.74 0.57***  Single mother family 0.20 0.18 0.33***  Other family type 0.08 0.08 0.10+ Family socioeconomic adversity (standardized) 0.05 −0.09 0.61*** Family receives food stamps 0.26 0.22 0.54*** Number of siblings in the household 1.46 1.42 1.75*** Length of time child has lived in current house (years) 2.77 2.81 2.49*** Child in fair/poor health 0.02 0.02 0.06*** Primary caregiver likely depressed 0.08 0.06 0.24*** Primary caregiver in fair/poor health 0.10 0.08 0.25*** Primary caregiver has health-related work limitations 0.07 0.06 0.16*** Panel B. Neighborhood Characteristics n = 3,800 Full Not Food Insecure Food Insecurea Neighborhood is urban 0.77 0.76 0.82*** Concentrated Disadvantage (standardized) 0.00 −0.13 0.33***  (components of concentrated disadvantage)   % Households in poverty 0.13 0.12 0.16***   % Female headed household 0.14 0.13 0.16***   % unemployed 0.09 0.08 0.10***   % Residents receiving public assistance 0.03 0.02 0.03*** Neighborhood Resource Measures (density per square mile)  Large supermarkets (50+ employees) 0.28 0.26 0.29*  Convenience Stores 0.46 0.39 0.53***  Social assistance organizations 2.89 2.47 3.13** Source: ECLS-K: 2010-2011 (N = 12,800 in 3,800 neighborhoods); 2007-2010 ACS data at the tract level a p-value for a t-test of significance by column (i.e. not food insecure vs. food insecure); *** ≤ 0.001 **≤ 0.01 *≤ 0.05 + ≤ 0.10 Table 1. Means and Proportions, ECLS-K: 2010-2011 Data Panel A. Individual and Family Characteristics n = 12,800 Full Sample Food Secure Food Insecurea Food insecure 0.12 Child is male 0.51 0.51 0.53 Child's age in months 74.60 74.50 74.80* Foreign-born parent 0.34 0.33 0.44*** Race/ethnicity  non-Hispanic white (ref) 0.51 0.54 0.34***  non-Hispanic black 0.11 0.10 0.15***  Hispanic 0.24 0.21 0.40***  Other 0.14 0.15 0.11*** Mother's age in years 34.50 34.70 33.20*** Family Structure  Two biological parents, married or cohab. (ref) 0.72 0.74 0.57***  Single mother family 0.20 0.18 0.33***  Other family type 0.08 0.08 0.10+ Family socioeconomic adversity (standardized) 0.05 −0.09 0.61*** Family receives food stamps 0.26 0.22 0.54*** Number of siblings in the household 1.46 1.42 1.75*** Length of time child has lived in current house (years) 2.77 2.81 2.49*** Child in fair/poor health 0.02 0.02 0.06*** Primary caregiver likely depressed 0.08 0.06 0.24*** Primary caregiver in fair/poor health 0.10 0.08 0.25*** Primary caregiver has health-related work limitations 0.07 0.06 0.16*** Panel B. Neighborhood Characteristics n = 3,800 Full Not Food Insecure Food Insecurea Neighborhood is urban 0.77 0.76 0.82*** Concentrated Disadvantage (standardized) 0.00 −0.13 0.33***  (components of concentrated disadvantage)   % Households in poverty 0.13 0.12 0.16***   % Female headed household 0.14 0.13 0.16***   % unemployed 0.09 0.08 0.10***   % Residents receiving public assistance 0.03 0.02 0.03*** Neighborhood Resource Measures (density per square mile)  Large supermarkets (50+ employees) 0.28 0.26 0.29*  Convenience Stores 0.46 0.39 0.53***  Social assistance organizations 2.89 2.47 3.13** Panel A. Individual and Family Characteristics n = 12,800 Full Sample Food Secure Food Insecurea Food insecure 0.12 Child is male 0.51 0.51 0.53 Child's age in months 74.60 74.50 74.80* Foreign-born parent 0.34 0.33 0.44*** Race/ethnicity  non-Hispanic white (ref) 0.51 0.54 0.34***  non-Hispanic black 0.11 0.10 0.15***  Hispanic 0.24 0.21 0.40***  Other 0.14 0.15 0.11*** Mother's age in years 34.50 34.70 33.20*** Family Structure  Two biological parents, married or cohab. (ref) 0.72 0.74 0.57***  Single mother family 0.20 0.18 0.33***  Other family type 0.08 0.08 0.10+ Family socioeconomic adversity (standardized) 0.05 −0.09 0.61*** Family receives food stamps 0.26 0.22 0.54*** Number of siblings in the household 1.46 1.42 1.75*** Length of time child has lived in current house (years) 2.77 2.81 2.49*** Child in fair/poor health 0.02 0.02 0.06*** Primary caregiver likely depressed 0.08 0.06 0.24*** Primary caregiver in fair/poor health 0.10 0.08 0.25*** Primary caregiver has health-related work limitations 0.07 0.06 0.16*** Panel B. Neighborhood Characteristics n = 3,800 Full Not Food Insecure Food Insecurea Neighborhood is urban 0.77 0.76 0.82*** Concentrated Disadvantage (standardized) 0.00 −0.13 0.33***  (components of concentrated disadvantage)   % Households in poverty 0.13 0.12 0.16***   % Female headed household 0.14 0.13 0.16***   % unemployed 0.09 0.08 0.10***   % Residents receiving public assistance 0.03 0.02 0.03*** Neighborhood Resource Measures (density per square mile)  Large supermarkets (50+ employees) 0.28 0.26 0.29*  Convenience Stores 0.46 0.39 0.53***  Social assistance organizations 2.89 2.47 3.13** Source: ECLS-K: 2010-2011 (N = 12,800 in 3,800 neighborhoods); 2007-2010 ACS data at the tract level a p-value for a t-test of significance by column (i.e. not food insecure vs. food insecure); *** ≤ 0.001 **≤ 0.01 *≤ 0.05 + ≤ 0.10 Approximately 12% of children in the ECLS-K Spring sample lived in households that met the criteria for food insecurity. Food insecurity in households with children is statistically similar for male and female children. Otherwise, important descriptive differences exist at the family level. Compared to children in food secure households, children in food insecure households are older, more likely to have a foreign-born parent, more likely black or Hispanic, have lived a shorter time at their current address, and have more siblings. More dramatic differences exist in the health status of the caregiver, in family structure, and in the socioeconomic situation of households. Roughly 6 and 8% of caregivers in food secure households are likely depressed and in fair or poor health, respectively. In food insecure households, a full 25% of caregivers are likely depressed and/or in fair or poor health. Further, about 18% of food secure households are single mother families whereas a third of all food insecure households are single mother families. Finally, food insecure households are, as expected, much more disadvantaged socioeconomically than are food secure households. For example, over half of food insecure households receive food stamps whereas less than a quarter of food secure households receive food stamps. Table 1 also shows that there are significant differences in the neighborhood characteristics of food insecure and food secure households. Children in food insecure households are more likely to live in more socially and economically disadvantaged neighborhoods. For example, children in food insecure households live in neighborhoods where on average 16% of all households are in poverty, compared to 12% for children in food secure households. Surprisingly, but consistent with other recent research (Sadler, Gilliland and Arku 2013), they are also located in neighborhoods with more supermarkets and with higher densities of convenience stores. This is partly a reflection of a larger proportion of ECLS-K children in urban than in rural neighborhoods overall. Indeed, 77% of children in the ECLS-K live in an urban neighborhood. Children in food insecure households are also more likely to live in neighborhoods with more social assistance organizations. To examine the relevance of neighborhood contributors for household food insecurity, we turn to the regression results in Table 2. Table 2. Multilevel Logistic Regression Models for Household Food Insecurity: Associations with Neighborhood Characteristics (Odds Ratios) Model 1 Model 2 Model 3 Model 4 Model 5 OR OR OR OR OR Child and Family Characteristics Child is male 1.04 1.04 1.04 1.04 1.04 Child's age in months 1.01 1.01 1.01 1.01 1.01 Foreign-born parent 1.26** 1.26** 1.27** 1.25** 1.26** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.86 0.85 0.86 0.85 0.86  Hispanic 1.06 1.05 1.07 1.06 1.06  Other 0.91 0.91 0.92 0.91 0.92 Mother's age in years 1.01* 1.01* 1.01* 1.01* 1.01* Family socioeconomic adversity (SEA) 2.33*** 2.32*** 2.33*** 2.33*** 2.33*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.14+ 1.15+ 1.15+ 1.15+  Other family type 0.82+ 0.82+ 0.82+ 0.82+ 0.82+ Number of siblings in the household 1.18*** 1.17*** 1.17*** 1.18*** 1.18*** Years lived in home 0.96* 0.96* 0.96* 0.96* 0.96* Family receives food stamps 1.66*** 1.66*** 1.66*** 1.66*** 1.66*** Child in fair/poor health 1.58** 1.58** 1.58** 1.58** 1.58** Primary caregiver likely depressed 2.89*** 2.89*** 2.89*** 2.89*** 2.89*** Primary caregiver in fair/poor health 1.58*** 1.57*** 1.58*** 1.58*** 1.58*** Primary caregiver has work limitations 1.54*** 1.54*** 1.54*** 1.54*** 1.54*** Neighborhood Characteristics Total Population (ln) 0.87* 0.86* 0.87+ 0.87* 0.87* Urban area 1.27** 1.27** 1.39*** 1.24* 1.31* Concentrated disadvantage 1.01 Neighborhood resources (density per square mile)  Large supermarkets (50+ employees) 0.97+  Convenience stores 1.01  Social assistance organizations 0.99 Constant 0.06*** 0.07*** 0.05*** 0.07*** 0.06*** sd(random intercept) 0.32*** 0.33*** 0.33*** 0.32*** 0.32*** sd(random slope - Family SEA) 0.14** 0.14** 0.14** 0.14** 0.14** Model 1 Model 2 Model 3 Model 4 Model 5 OR OR OR OR OR Child and Family Characteristics Child is male 1.04 1.04 1.04 1.04 1.04 Child's age in months 1.01 1.01 1.01 1.01 1.01 Foreign-born parent 1.26** 1.26** 1.27** 1.25** 1.26** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.86 0.85 0.86 0.85 0.86  Hispanic 1.06 1.05 1.07 1.06 1.06  Other 0.91 0.91 0.92 0.91 0.92 Mother's age in years 1.01* 1.01* 1.01* 1.01* 1.01* Family socioeconomic adversity (SEA) 2.33*** 2.32*** 2.33*** 2.33*** 2.33*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.14+ 1.15+ 1.15+ 1.15+  Other family type 0.82+ 0.82+ 0.82+ 0.82+ 0.82+ Number of siblings in the household 1.18*** 1.17*** 1.17*** 1.18*** 1.18*** Years lived in home 0.96* 0.96* 0.96* 0.96* 0.96* Family receives food stamps 1.66*** 1.66*** 1.66*** 1.66*** 1.66*** Child in fair/poor health 1.58** 1.58** 1.58** 1.58** 1.58** Primary caregiver likely depressed 2.89*** 2.89*** 2.89*** 2.89*** 2.89*** Primary caregiver in fair/poor health 1.58*** 1.57*** 1.58*** 1.58*** 1.58*** Primary caregiver has work limitations 1.54*** 1.54*** 1.54*** 1.54*** 1.54*** Neighborhood Characteristics Total Population (ln) 0.87* 0.86* 0.87+ 0.87* 0.87* Urban area 1.27** 1.27** 1.39*** 1.24* 1.31* Concentrated disadvantage 1.01 Neighborhood resources (density per square mile)  Large supermarkets (50+ employees) 0.97+  Convenience stores 1.01  Social assistance organizations 0.99 Constant 0.06*** 0.07*** 0.05*** 0.07*** 0.06*** sd(random intercept) 0.32*** 0.33*** 0.33*** 0.32*** 0.32*** sd(random slope - Family SEA) 0.14** 0.14** 0.14** 0.14** 0.14** *** < 0.001 ** < 0.01 * < 0.05 + < 0.10 Table 2. Multilevel Logistic Regression Models for Household Food Insecurity: Associations with Neighborhood Characteristics (Odds Ratios) Model 1 Model 2 Model 3 Model 4 Model 5 OR OR OR OR OR Child and Family Characteristics Child is male 1.04 1.04 1.04 1.04 1.04 Child's age in months 1.01 1.01 1.01 1.01 1.01 Foreign-born parent 1.26** 1.26** 1.27** 1.25** 1.26** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.86 0.85 0.86 0.85 0.86  Hispanic 1.06 1.05 1.07 1.06 1.06  Other 0.91 0.91 0.92 0.91 0.92 Mother's age in years 1.01* 1.01* 1.01* 1.01* 1.01* Family socioeconomic adversity (SEA) 2.33*** 2.32*** 2.33*** 2.33*** 2.33*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.14+ 1.15+ 1.15+ 1.15+  Other family type 0.82+ 0.82+ 0.82+ 0.82+ 0.82+ Number of siblings in the household 1.18*** 1.17*** 1.17*** 1.18*** 1.18*** Years lived in home 0.96* 0.96* 0.96* 0.96* 0.96* Family receives food stamps 1.66*** 1.66*** 1.66*** 1.66*** 1.66*** Child in fair/poor health 1.58** 1.58** 1.58** 1.58** 1.58** Primary caregiver likely depressed 2.89*** 2.89*** 2.89*** 2.89*** 2.89*** Primary caregiver in fair/poor health 1.58*** 1.57*** 1.58*** 1.58*** 1.58*** Primary caregiver has work limitations 1.54*** 1.54*** 1.54*** 1.54*** 1.54*** Neighborhood Characteristics Total Population (ln) 0.87* 0.86* 0.87+ 0.87* 0.87* Urban area 1.27** 1.27** 1.39*** 1.24* 1.31* Concentrated disadvantage 1.01 Neighborhood resources (density per square mile)  Large supermarkets (50+ employees) 0.97+  Convenience stores 1.01  Social assistance organizations 0.99 Constant 0.06*** 0.07*** 0.05*** 0.07*** 0.06*** sd(random intercept) 0.32*** 0.33*** 0.33*** 0.32*** 0.32*** sd(random slope - Family SEA) 0.14** 0.14** 0.14** 0.14** 0.14** Model 1 Model 2 Model 3 Model 4 Model 5 OR OR OR OR OR Child and Family Characteristics Child is male 1.04 1.04 1.04 1.04 1.04 Child's age in months 1.01 1.01 1.01 1.01 1.01 Foreign-born parent 1.26** 1.26** 1.27** 1.25** 1.26** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.86 0.85 0.86 0.85 0.86  Hispanic 1.06 1.05 1.07 1.06 1.06  Other 0.91 0.91 0.92 0.91 0.92 Mother's age in years 1.01* 1.01* 1.01* 1.01* 1.01* Family socioeconomic adversity (SEA) 2.33*** 2.32*** 2.33*** 2.33*** 2.33*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.14+ 1.15+ 1.15+ 1.15+  Other family type 0.82+ 0.82+ 0.82+ 0.82+ 0.82+ Number of siblings in the household 1.18*** 1.17*** 1.17*** 1.18*** 1.18*** Years lived in home 0.96* 0.96* 0.96* 0.96* 0.96* Family receives food stamps 1.66*** 1.66*** 1.66*** 1.66*** 1.66*** Child in fair/poor health 1.58** 1.58** 1.58** 1.58** 1.58** Primary caregiver likely depressed 2.89*** 2.89*** 2.89*** 2.89*** 2.89*** Primary caregiver in fair/poor health 1.58*** 1.57*** 1.58*** 1.58*** 1.58*** Primary caregiver has work limitations 1.54*** 1.54*** 1.54*** 1.54*** 1.54*** Neighborhood Characteristics Total Population (ln) 0.87* 0.86* 0.87+ 0.87* 0.87* Urban area 1.27** 1.27** 1.39*** 1.24* 1.31* Concentrated disadvantage 1.01 Neighborhood resources (density per square mile)  Large supermarkets (50+ employees) 0.97+  Convenience stores 1.01  Social assistance organizations 0.99 Constant 0.06*** 0.07*** 0.05*** 0.07*** 0.06*** sd(random intercept) 0.32*** 0.33*** 0.33*** 0.32*** 0.32*** sd(random slope - Family SEA) 0.14** 0.14** 0.14** 0.14** 0.14** *** < 0.001 ** < 0.01 * < 0.05 + < 0.10 The results for Model 1 of Table 2 are largely in line with the body of empirical research on food insecurity, showing that characteristics of children and their families are strongly associated with the odds of household food insecurity. That is, children who are older, have a foreign-born parent (relative to those without), have older mothers, and have primary caregivers who report health and disability issues have higher odds of food insecurity. In particular, if the caregiver is likely depressed, the household in which the child lives has nearly three times higher odds of food insecurity, relative to a household where the caregiver is not likely depressed. Finally, families at greater socioeconomic adversity face significantly higher odds of food insecurity. Controlling for all other child and family characteristics, a one standard deviation increase in socioeconomic adversity is associated with 2.3 times higher odds of household food insecurity. Models 2-5 show that these child and family associations with household food insecurity are robust and go largely unchanged as neighborhood characteristics enter the model. More important for our purposes, we find no significant associations with food insecurity and neighborhood concentrated disadvantage (Model 2), density of supermarkets or convenience stores (Models 3 and 4), or social assistance organizations (Model 5). Thus, we find no support for hypotheses 1 to 3 – neighborhood characteristics are not associated with food insecurity risk, accounting for important child and family characteristics. The random components at the bottom of the table suggest that the odds of food insecurity do indeed vary across neighborhoods (random intercept) and, perhaps more importantly, the effect of family socioeconomic adversity varies across neighborhoods (random slope). Thus, we turn to asking whether neighborhood disadvantage moderates the impact of family socioeconomic adversity on food insecurity. Table 3 provides results for cross-level interactions that allow us to evaluate whether children in the most disadvantaged families face special heightened risks of food insecurity when they live in the most disadvantaged neighborhoods (Hypothesis 4) or whether a disadvantage paradox exists and these children actually have a lower risk of food insecurity (Hypothesis 5). Specifically, we provide results for an interaction between family socioeconomic adversity and neighborhood concentrated disadvantage. We also tested interactions between family socioeconomic adversity and the neighborhood density of large supermarkets, convenience stores, and social assistance organizations, but they were not significant so we do not present them here (full results available upon request). Table 3. Multilevel Logistic Regression Models for Household Food Insecurity: Interaction Models (Odds Ratios) Model 1 OR Model 2 OR Child and Family Characteristics Child is male 1.04 1.04 Child's age in months 1.01 1.01 Foreign-born parent 1.27** 1.28** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.82+ 0.82+  Hispanic 1.02 1.02  Other 0.88 0.89 Mother's age in years 1.01** 1.01** Family socioeconomic adversity (SEA) 2.35*** 2.35*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.15+  Other family type 0.79* 0.79* Number of siblings in the household 1.18*** 1.18*** Years lived in home 0.96* 0.96* Family receives food stamps 1.63*** 1.63*** Child in fair/poor health 1.59** 1.59** Primary caregiver likely depressed 2.90*** 2.90*** Primary caregiver in fair/poor health 1.56*** 1.56*** Primary caregiver has work limitations 1.52*** 1.52*** Neighborhood Characteristics Total Population (ln) 0.87* 0.87* Urban area 1.29** 1.33* Concentrated disadvantage 1.22*** 1.22*** Concentrated disadvantage*Family SEA 0.76*** 0.76*** Neighborhood resources (density per square mile) Social assistance organizations 0.99 Constant 0.06*** 0.06*** sd(random intercept) 0.30*** 0.30*** sd(random slope - Family SEA) 0.11* 0.11* Model 1 OR Model 2 OR Child and Family Characteristics Child is male 1.04 1.04 Child's age in months 1.01 1.01 Foreign-born parent 1.27** 1.28** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.82+ 0.82+  Hispanic 1.02 1.02  Other 0.88 0.89 Mother's age in years 1.01** 1.01** Family socioeconomic adversity (SEA) 2.35*** 2.35*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.15+  Other family type 0.79* 0.79* Number of siblings in the household 1.18*** 1.18*** Years lived in home 0.96* 0.96* Family receives food stamps 1.63*** 1.63*** Child in fair/poor health 1.59** 1.59** Primary caregiver likely depressed 2.90*** 2.90*** Primary caregiver in fair/poor health 1.56*** 1.56*** Primary caregiver has work limitations 1.52*** 1.52*** Neighborhood Characteristics Total Population (ln) 0.87* 0.87* Urban area 1.29** 1.33* Concentrated disadvantage 1.22*** 1.22*** Concentrated disadvantage*Family SEA 0.76*** 0.76*** Neighborhood resources (density per square mile) Social assistance organizations 0.99 Constant 0.06*** 0.06*** sd(random intercept) 0.30*** 0.30*** sd(random slope - Family SEA) 0.11* 0.11* *** < 0.001 ** < 0.01 * < 0.05 + < 0.10 Table 3. Multilevel Logistic Regression Models for Household Food Insecurity: Interaction Models (Odds Ratios) Model 1 OR Model 2 OR Child and Family Characteristics Child is male 1.04 1.04 Child's age in months 1.01 1.01 Foreign-born parent 1.27** 1.28** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.82+ 0.82+  Hispanic 1.02 1.02  Other 0.88 0.89 Mother's age in years 1.01** 1.01** Family socioeconomic adversity (SEA) 2.35*** 2.35*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.15+  Other family type 0.79* 0.79* Number of siblings in the household 1.18*** 1.18*** Years lived in home 0.96* 0.96* Family receives food stamps 1.63*** 1.63*** Child in fair/poor health 1.59** 1.59** Primary caregiver likely depressed 2.90*** 2.90*** Primary caregiver in fair/poor health 1.56*** 1.56*** Primary caregiver has work limitations 1.52*** 1.52*** Neighborhood Characteristics Total Population (ln) 0.87* 0.87* Urban area 1.29** 1.33* Concentrated disadvantage 1.22*** 1.22*** Concentrated disadvantage*Family SEA 0.76*** 0.76*** Neighborhood resources (density per square mile) Social assistance organizations 0.99 Constant 0.06*** 0.06*** sd(random intercept) 0.30*** 0.30*** sd(random slope - Family SEA) 0.11* 0.11* Model 1 OR Model 2 OR Child and Family Characteristics Child is male 1.04 1.04 Child's age in months 1.01 1.01 Foreign-born parent 1.27** 1.28** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.82+ 0.82+  Hispanic 1.02 1.02  Other 0.88 0.89 Mother's age in years 1.01** 1.01** Family socioeconomic adversity (SEA) 2.35*** 2.35*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.15+  Other family type 0.79* 0.79* Number of siblings in the household 1.18*** 1.18*** Years lived in home 0.96* 0.96* Family receives food stamps 1.63*** 1.63*** Child in fair/poor health 1.59** 1.59** Primary caregiver likely depressed 2.90*** 2.90*** Primary caregiver in fair/poor health 1.56*** 1.56*** Primary caregiver has work limitations 1.52*** 1.52*** Neighborhood Characteristics Total Population (ln) 0.87* 0.87* Urban area 1.29** 1.33* Concentrated disadvantage 1.22*** 1.22*** Concentrated disadvantage*Family SEA 0.76*** 0.76*** Neighborhood resources (density per square mile) Social assistance organizations 0.99 Constant 0.06*** 0.06*** sd(random intercept) 0.30*** 0.30*** sd(random slope - Family SEA) 0.11* 0.11* *** < 0.001 ** < 0.01 * < 0.05 + < 0.10 Turning to Model 1, we see that neighborhood concentrated disadvantage interacts with family socioeconomic adversity such that neighborhood disadvantage is associated with 22% higher odds of food insecurity for families at the mean level of socioeconomic adversity, but that this association diminishes as family socioeconomic adversity increases. In other words, families with more socioeconomic adversity have lower odds of food insecurity as neighborhood disadvantage increases. To illustrate, we generated predicted probabilities for food insecurity for three categories of family socioeconomic adversity: Mean Family socioeconomic adversity, Higher Family socioeconomic adversity (families at two standard deviations above the mean), and Lower Family socioeconomic adversity (families at two standard deviations below the mean). In Figure 1, we plot probabilities of food insecurity for the three groups by standardized neighborhood concentrated disadvantage. So for example, when neighborhood concentrated disadvantage equals zero (or is at the mean), families with low socioeconomic adversity have a probability of food insecurity of .03; families at the mean level of socioeconomic adversity have a probability of .10; and those with high socioeconomic adversity have a probability of .36. In general, the figure shows that the probability of food insecurity increases with neighborhood disadvantage for children in families at the mean level of socioeconomic adversity or lower.2 But for children in families at high levels of socioeconomic adversity, the probability of food insecurity dramatically decreases as neighborhood disadvantage increases, providing support for a disadvantage paradox hypothesis. Figure 1. View largeDownload slide Predicted Probability of Household Food Insecurity by Family Socioeconomic Adversity and Neighborhood Concentrated Disadvantage Figure 1. View largeDownload slide Predicted Probability of Household Food Insecurity by Family Socioeconomic Adversity and Neighborhood Concentrated Disadvantage Finally, in Model 2 of Table 3 we examined whether the cross-level interaction between neighborhood concentrated disadvantage and family socioeconomic adversity is robust after including the density of social assistance organizations. In other words, is lower risk for food insecurity for low SES families in higher disadvantaged neighborhoods due to greater access to social services in those neighborhoods? Accounting for social service organizations in Model 2 does not change the relationship between family socioeconomic adversity and neighborhood concentrated disadvantage. Put differently, this paradoxical relationship is not explained by accounting for a higher density of social assistance organizations in the neighborhood. DISCUSSION Scholars interested in poverty, collective action, and how communities impact individuals should be concerned with household food insecurity. With over 20% of households with children in the U.S. struggling to acquire enough food for a healthy and active lifestyle and signs that this problem is on the rise rather than receding, researchers, child advocates, policy makers, and the nation as a whole have a responsibility to better understand and provide solutions to curb and ultimately eliminate food insecurity. Our results demonstrate that food insecurity is not a simple or predictable measure of household disadvantage. Rather, food insecurity may be unevenly distributed even among the disadvantaged. Using nationally representative data, we clarify if and to what extent neighborhood characteristics may be influencing food insecurity in households that include young children. First, we investigated whether neighborhood characteristics predicted household food insecurity over and above individual and/or family characteristics. Similar to other recent studies (Kirkpatrick and Tarasuk 2010; Sadler, Gilliland and Arku 2013), we find that family, not neighborhood, characteristics are most directly linked with food insecurity. Indeed, families with young children in the U.S. facing social and economic disadvantages are struggling to provide enough food for their households. Alleviating poverty, ensuring equal access to educational and employment opportunities, and similar broad sweeping policy goals, can provide far reaching benefits for families with young children. But not all households in poverty are food insecure, and some non-poor households struggle with food security (Gundersen, Kreider and Pepper 2011). Thus, we posited that community characteristics might impact well-being differently for families at different levels of disadvantage. Indeed, existing evidence suggests that the impact of neighborhood disadvantage (Wodtke, Harding and Elwert 2016) and the ways in which communities might respond to adversity (Crowe and Smith 2012) are not monolithic. We presented and tested hypotheses on competing perspectives aimed at understanding how neighborhood and family characteristics impacted children in the most disadvantaged families – would family and neighborhood disadvantage accumulate to increase the risk of food insecurity, or would they interact in unexpected ways? Our results shed new light on the apparent null relationship between neighborhood disadvantage and household food insecurity. Increasing neighborhood disadvantage does associate with increasing risk of food insecurity for more well-off families. But for the most disadvantaged families, we find support for a disadvantage paradox – children in the most disadvantaged families and living in the most disadvantaged neighborhoods have lower probabilities of food insecurity than similar children in less disadvantaged neighborhoods. We investigated the possibility that this was a result of access to greater social resources, such as the placement of food pantries and other social organizations in the most disadvantaged neighborhoods, and find that these resources do little to explain the paradox. Rather, although we are unable to test this mechanism with our data, this finding may be a result of vulnerable residents taking on more informal strategies to deal with their own disadvantaged circumstances (Desmond 2012; Klinenberg 2002; Stack 1974). In all, these results suggest a contextual incongruence between family and neighborhood characteristics impacting the risk of food insecurity. In other words, when the level of family disadvantage is substantively different from the level of neighborhood disadvantage, resources that may be leveraged to mitigate food insecurity may be more difficult to access. Further, the neighborhood results suggest that social processes, and not necessarily physical access to food, are at the core of understanding how environment influences risk of food insecurity. Our null results regarding access to grocery stores are consistent with a growing body of work that fails to find robust associations between physical access to quality foods and well-being (Cummins, Flint and Matthews 2014; Sadler, Gilliland and Arku 2013). In addition, we find little evidence that increased access to formal social services influences risk, suggesting that families in need may engage in more informal ways to share information and obtain sufficient food supplies. Poor families in the most destitute communities may be employing effective strategies to share information and resources to counter food insecurity, particularly when it impacts children (Morton et al. 2005; Piff et al. 2010; Stack 1974). For example, dense, reciprocal ties among single mothers may be an important survival strategy in low-income communities (Dominguez and Watkins 2003), and participation in, not just the presence of, social organizations may be another survival strategy (Rankin and Quane 2000). This suggests that there may be close links among the poor which resemble those in mid-century American suburbs (Newman 1992), and that the meaning and use of social capital differs in high versus low SES neighborhoods (Altschuler, Somkin and Adler 2004; Sherman 2006). For example, residents of high SES neighborhoods may utilize social capital more to gain access to resources like job leads, but residents of low SES neighborhoods may use those ties more for daily survival (Desmond 2012). These social connections may be strongest among the most disadvantaged families in the most disadvantaged neighborhoods, possibly a product of moral capital and a closely shared set of strategies on how to cope with extreme adversity (Sherman 2006). Though not as often a research focus, scholars have documented coordinated strategies among residents in disadvantaged communities to overcome seemingly overwhelming hardship (Seccombe 2002). Indeed, prior evidence and the findings presented here support the idea that unique social processes are at work in disadvantaged communities. But more work is needed. Specifically, how might resource-sharing in poor communities work to reduce food insecurity and what sorts of policies can support those efforts? As a start, we might turn to existing policies, such as free and reduced breakfast and lunch programs in schools, and expand similar programs through community centers, especially when school is out of session. Such community food centers could purchase food in bulk, reducing cost, and provide education and training programs targeted to specific nutritional goals matched to the kinds of foods available. In addition, after-school programs which feed children (and sometimes their families) dinner could be expanded at schools with a higher proportion of students who receive free and reduced lunch. Such programs harness and nurture social capital and can address food insecurity at a community level. That children in less disadvantaged families in more disadvantaged neighborhoods do not receive the same reduction to food insecurity risk as children in the most disadvantaged families is interesting. In lower-income communities, social resources are mobilized to address chronic problems, and to gain access to resources residents cannot through financial means (Altschuler, Somkin and Adler 2004). These reciprocal exchanges of goods and services are strongest among residents of similar means (Dominguez and Watkins 2003; Menjivar 2000). Given these findings in conjunction with our own, we believe this is compelling evidence that less disadvantaged families may be more isolated in more disadvantaged communities, and less able (or less willing) to access the social networks of their more disadvantaged neighbors or social organizations which are designed to provide assistance. It is important to note that the ECLS-K is weighted toward children living in urban areas, and so some of these processes may differ for households with children in rural areas. Gaining a better understanding of the mechanisms underlying this “disadvantage paradox” will not only illuminate how community resources can be leveraged to alleviate food insecurity, but also how social ties and community resources are utilized by the poor to survive. The first and second authors acknowledge support from the University of Kentucky Center for Poverty Research (UKCPR) through funding from the Food and Nutrition Service of the Department of Agriculture (Contract No. AG-3198-S-12-0044) and the first author acknowledges support from the Young Scholars Program of the Foundation for Child Development (FCD) (Grant No. YSP Rice 10-2014). The opinions and conclusions expressed herein are solely those of the authors and should not be construed as representing the opinions or policies of the UKCPR, the FCD, or any agency of the Federal Government. The authors thank Laura Freeman for research assistance and the Kinder Institute Urban Health Program at Rice University for administrative support. Footnotes 1 Zip codes were converted to census tracts via crosswalk files generated from the Missouri Census Data Center’s MABLE/Geocorr2K application (http://mcdc2.missouri.edu/websas/geocorr2k.html). 2 Because the number of poor households in advantaged neighborhoods and the number of wealthy households in disadvantaged neighborhoods is relatively small, we re-estimated our models after eliminating outliers and found much the same results (available upon request). For presentation, the figure represents condensed results after eliminating outliers. REFERENCES Center for Law and Public Policy, (CLASP) . 2013 . Child Poverty in the U.S.: What New Census Data Tell Us About Our Youngest Children. Washington DC . Ahluwalia Indu B. , Dodds Janice M. , Baligh Magda . 1998 . “Social support and coping behaviors of low-income families experiencing food insufficiency in North Carolina.” Health Education & Behavior 25 5 : 599 - 612 . Alaimo K. , Briefel R. R. , Frongillo E. A. , Olson C. M. . 1998 . “Food insufficiency exists in the United States: Results from the third National Health and Nutrition Examination Survey (NHANES III).” American Journal of Public Health 88 3 : 419 - 26 . Alaimo K. , Olson C. M. , Frongillo E. A. . 2002 . “Family food insufficiency, but not low family income, is positively associated with dysthymia and suicide symptoms in adolescents.” J Nutr 132 4 : 719 - 25 . Alaimo K. , Olson C. M. , Frongillo E. A. Jr. 2001 . “Food insufficiency and American school-aged children's cognitive, academic, and psychosocial development.” Pediatrics 108 1 : 44 - 53 . Alaimo K. , Olson C. M. , Frongillo E. A. Jr. , Briefel R. R. . 2001 . “Food insufficiency, family income, and health in US preschool and school-aged children.” Am J Public Health 91 5 : 781 - 6 . Allison P.D. 2002 . Missing Data : London : Sage University . Altschuler Andrea , Somkin Carol P. , Adler Nancy E. . 2004 . “Local services and amenities, neighborhood social capital, and health.” Social Science & Medicine 59 6 : 1219 - 29 . Anderson Elijah. 1990 . Streetwise: Race, Class, and Change in an Urban Community . Chicago, IL : University of Chicago Press . Anderson Elijah. 1999 . Code of the Street: Decency, Violence, and the Moral Life of the Inner City . New York : W.W. Norton & Company . Bell Bethany A. , Ferron John M. , Kromrey Jeffrey D. . 2008 . “Cluster Size in Multilevel Models: The Impact of Sparse Data Structures on Point and Interval Estimates in Two-Level Models.” in Proceedings of the American Statistical Association, Section on Survey Research Methods . Alexandria, VA : American Statistical Association . Bickel Gary , Nord Mark , Price Cristofer , Hamilton William , Cook John . 2000 . “Guide to Measuring Household food security, revised 2000.” Alexandria, VA : United States Department of Agriculture, Food and Nutrition Service . Borman Geoffrey D. , Dowling Maritza . 2010 . “Schools and Inequality: A Multilevel Analysis of Coleman's Equality of Educational Opportunity Data ” Teachers College Record 112 5 : 1201 - 46 . Carpiano R.M. , Lloyd J.E.V. , Hertzman C. . 2009 . “Concentrated affluence, concentrated disadvantage, and children's readiness for school: A population-based, multi-level investigation.” Social Science & Medicine 69 3 : 420 - 32 . Carpiano Richard M. 2008 . “Actual or potential neighborhood resources and access to them: Testing hypotheses of social capital for the health of female caregivers.” Social Science & Medicine 67 4 : 568 - 82 . Casey P. H. , Szeto K. , Lensing S. , Bogle M. , Weber J. . 2001 . “Children in food-insufficient, low-income families: prevalence, health, and nutrition status.” Arch Pediatr Adolesc Med 155 4 : 508 - 14 . Coleman-Jensen Alisha , Gregory Christian , Singh Anita . 2014 . “Household Food Security in the United States in 2013: Statistical Supplement.” Economic Research Service U.S. Department of Agriculture . Cook J. T. , Frank D. A. , Berkowitz C. , Black M. M. , Casey P. H. , Cutts D. B. , Meyers A. F. , Zaldivar N. , Skalicky A. , Levenson S. , Heeren T. , Nord M. . 2004 . “Food insecurity is associated with adverse health outcomes among human infants and toddlers.” J Nutr 134 6 : 1432 - 8 . Crowe Jessica , Smith Justin . 2012 . “The influence of community capital toward a community's capacity to respond to food insecurity.” Community Development 43 2 : 169 - 86 . Cummins Steven , Flint Ellen , Matthews Stephen A. . 2014 . “New Neighborhood Grocery Store Increased Awareness Of Food Access But Did Not Alter Dietary Habits Or Obesity.” Health Affairs 33 2 : 283 - 91 . De Marco M. , Thorburn S. . 2009 . “The relationship between income and food insecurity among Oregon residents: does social support matter?” Public Health Nutr 12 11 : 2104 - 12 . Desmond Matthew. 2012 . “Disposable Ties and the Urban Poor.” American Journal of Sociology 117 5 : 1295 - 335 . Dominguez Silvia , Watkins Celeste . 2003 . “Creating networks for survival and mobility: Social capital among African-American and Latin-American low-income mothers.” Social Problems 50 1 : 111 - 35 . Fram Maryah Stella , Frongillo Edward A. , Draper Carrie L. , Fishbein Eliza M. . 2013 . “Development and Validation of a Child Report Assessment of Child Food Insecurity and Comparison to Parent Report Assessment.” Journal of Hunger & Environmental Nutrition 8 2 : 128 - 45 . Fram Maryah Stella , Frongillo Edward A. , Jones Sonya J. , Williams Roger C. , Burke Michael P. , DeLoach Kendra P. , Blake Christine E. . 2011 . “Children are aware of food insecurity and take responsibility for managing food resources.” The Journal of Nutrition 141 6 : 1114 - 19 . Gundersen C. , Kreider B. , Pepper J. . 2011 . “The Economics of Food Insecurity in the United States.” Applied Economic Perspectives and Policy 33 3 : 281 - 303 . Guo Guang , Zhao Hongxin . 2000 . “Multilevel Modeling for Binary Data.” Annual Review of Sociology 26 : 441 - 62 . Halpern-Meekin Sarah , Edin Kathryn , Tach Laura , Sykes Jennifer . 2015 . It's Not Like I'm Poor: How Working Families Make Ends Meet in a Post-welfare World . Oakland, CA : Univ of California Press . Hung Chanjin , Myers Samuel L. . 1999 . “Do the poor pay more for food? An analysis of grocery store availability and food price disparities.” The Journal of Consumer Affairs 33 2 : 276 - 96 . Jencks Christopher , Mayer Susan E. . 1990 . “The Social Consequences of Growing Up in a Poor Neighborhood.” Pp. 111 - 85 in Inner-City Poverty in the United States , edited by Lynn Laurence E. , McGeary Michael G. H. . Washington, D.C .: National Academy Press . Jetter Karen M. , Cassady Diana L. . 2006 . “The availability and cost of healthier food alternatives.” American Journal of Preventive Medicine 30 1 : 38 - 44 . Jones S. J. , Jahns L. , Laraia B. A. , Haughton B. . 2003 . “Lower risk of overweight in school-aged food insecure girls who participate in food assistance: results from the panel study of income dynamics child development supplement.” Arch Pediatr Adolesc Med 157 8 : 780 - 4 . Jyoti D. F. , Frongillo E. A. , Jones S. J. . 2005 . “Food insecurity affects school children's academic performance, weight gain, and social skills.” Journal of Nutrition 135 12 : 2831 - 39 . Kalil A. , Chen J. H. . 2008 . “Mothers' citizenship status and household food insecurity among low-income children of immigrants.” New Dir Child Adolesc Dev 2008 121 : 43 - 62 . Kawachi Ichiro. 2010 . “Social Capital and Health.” Pp. 18 - 32 in Handbook of Medical Sociology , edited by Bird C.E. , Conrad P. , Fremont A.M. , Timmermans Stefan . Nashville, TN : Vanderbilt University Press . Kawachi Ichiro , Berkman Lisa F. (Eds.). 2003 . Neighborhoods and Health . New York : Oxford University Press . Kimbro Rachel Tolbert , Denney Justin T . 2015 . “Transitions Into Food Insecurity Associated With Behavioral Problems And Worse Overall Health Among Children.” Health Affairs 34 11 : 1949 - 55 . Kimbro Rachel Tolbert , Denney Justin T. , Panchang Sarita . 2012 . “Individual, Family, and Neighborhood Characteristics and Children’s Food Insecurity.” Journal of Applied Research on Children 3 1 :Article 8. Kirkpatrick S.I. , McIntyre L. , Potestio M.L. . 2010 . “Child hunger and long-term adverse consequences for health.” Archives of Pediatrics and Adolescent Medicine 164 8 : 754 . Kirkpatrick Sharon I. , Tarasuk Valerie . 2010 . “Assessing the Relevance of Neighbourhood Characteristics to the Household Food Security of Low-Income Toronto Families.” Public Health Nutrition 13 7 : 1139 - 48 . Kirkpatrick Sharon I. , Tarasuk Valerie 2011 . “Housing Circumstances are Associated with Household Food Access Among Low-Income Urban Families.” Journal of Urban Health 88 2 : 284 - 96 . Kleinman R. E. , Murphy J. M. , Little M. , Pagano M. , Wehler C. A. , Regal K. , Jellinek M. S. . 1998 . “Hunger in children in the United States: potential behavioral and emotional correlates.” Pediatrics 101 1 : E3 . Klinenberg Eric. 2002 . Heat Wave: A Social Autopsy of Disaster in Chicago . Chicago, IL : The University of Chicago Press . Leventhal T. , Brooks-Gunn J. . 2000 . “The neighborhoods they live in: The effects of neighborhood residence on child and adolescent outcomes.” Psychological Bulletin 126 2 : 309 - 37 . Lichter Daniel T , Brown David L . 2011 . “Rural America in an urban society: Changing spatial and social boundaries.” Annual Review of Sociology 37 : 565 - 92 . Martin Katie S. , Rogers Beatrice L. , Cook John T. , Joseph Hugh M. . 2004 . “Social capital is associated with decreased risk of hunger.” Social Science & Medicine 58 12 : 2645 - 54 . McIntyre Lynn , Williams Jeanne V. A. , Lavorato Dina H. , Patten Scott . 2013 . “Depression and suicide ideation in late adolescence and early adulthood are an outcome of child hunger.” Journal of Affective Disorders 150 1 : 123 - 29 . Menjivar Cecilia. 2000 . Fragmented ties: Salvadoran immigrant networks in America . Berkeley, CA : University of California Press . Morenoff J. D. 2003 . “Neighborhood mechanisms and the spatial dynamics of birth weight.” American Journal of Sociology 108 5 : 976 - 1017 . Morland Kimberly , Diez Roux Ana V. , Wing Steve . 2006 . “Supermarkets, other food stores, and obesity: the atherosclerosis risk in communities study.” American Journal of Preventive Medicine 30 4 : 333 - 39 . Morton Lois Wright , Bitto Ella Annette , Oakland Mary Jane , Sand Mary . 2005 . “Solving the Problems of Iowa Food Deserts: Food Insecurity and Civic Structure.” Rural Sociology 70 1 : 94 - 112 . Morton Lois Wright , Bitto Ella Annette , Oakland Mary Jane , Sand Mary 2008 . “Accessing food resources: Rural and urban patterns of giving and getting food.” Agriculture and Human Values 25 1 : 107 - 19 . Murphy J. M. , Wehler C. A. , Pagano M. E. , Little M. , Kleinman R. E. , Jellinek M. S. . 1998 . “Relationship between hunger and psychosocial functioning in low-income American children.” J Am Acad Child Adolesc Psychiatry 37 2 : 163 - 70 . Newman Katherine S. 1992 . “Culture and structure in the truly disadvantaged.” City & Society 6 1 : 3 - 25 . Nord Mark , Coleman-Jensen Alisha , Andrews Margaret , Carlson Steven . 2010 . “Household food security in the United States, 2009.” Economic Research Report . Washington D.C .: US Department of Agriculture . Nord Mark , Hanson Karla . 2012 . “Adult caregiver reports of adolescents' food security do not agree well with adolescents' own reports.” Journal of Hunger & Environmental Nutrition 7 4 : 363 - 80 . Piff P.K. , Kraus M.W. , Côté S. , Cheng B.H. , Keltner D. . 2010 . “Having less, giving more: The influence of social class on prosocial behavior.” Journal of Personality and Social Psychology 99 5 : 771 . Rabe-Hesketh Sophia, , Skrondal Anders . 2008 . Multilevel and Longitudinal Modeling Using Stata . College Station, TX : StataCorp . Radloff Lenore Sawyer. 1977 . “The CES-D Scale: A Self-Report Depression Scale for Research in the General Population.” Applied Psychological Measurement 1 : 385 - 401 . Rankin Bruce H. , Quane James M. . 2000 . “Neighborhood poverty and the social isolation of inner-city African American families.” Social Forces 79 1 : 139 - 64 . Raudenbush Stephen W. , Bryk Anthony S. . 2002 . Hierarchical Linear Models . Thousand Oaks, CA : Sage Publications . Rose D. , Richards R. . 2004 . “Food store access and household fruit and vegetable use among participants in the US Food Stamp Program.” Public Health Nutrition 7 8 : 1081 - 88 . Royston Patrick. 2005 . “Multiple imputation of missing values: update.” The Stata Journal 5 2 : 188 - 201 . Ryu Jeong-Hee , Bartfeld Judith S. . 2012 . “Household food insecurity during childhood and subsequent health status: the Early Childhood Longitudinal Study--Kindergarten Cohort.” American Journal of Public Health 102 11 : e50 - e55 . Sadler Richard C. , Gilliland Jason A. , Arku Godwin . 2013 . “A Food Retail-Based Intervention on Food Security and Consumption.” International journal of environmental research and public health 10 8 : 3325 - 46 . Sampson Robert J. 2012 . Great American city: Chicago and the enduring neighborhood effect . Chicago, IL : University of Chicago Press . Sampson Robert J. , Raudenbush Stephen W. , Earls Felton . 1997 . “Neighborhoods and Violent Crime: A Multilevel Study of Collective Efficacy.” Science 277 5328 : 918 - 24 . Seccombe Karen. 2002 . ““Beating the odds” versus “changing the odds”: Poverty, resilience, and family policy.” Journal of Marriage and Family 64 2 : 384 - 94 . Sherman Jennifer. 2006 . “Coping with rural poverty: Economic survival and moral capital in rural America.” Social Forces 85 2 : 891 - 913 . Small Mario Luis. 2004 . Villa Victoria: The transformation of social capital in a Boston barrio : University of Chicago Press . Small Mario Luis. 2009 . Unanticipated gains: Origins of network inequality in everyday life . New York, NY : Oxford University Press . Small Mario Luis , Harding David J. , Lamont Michele . 2010 . “Reconsidering culture and poverty.” ANNALS, AAPSS 629 : 6 - 26 . Smith Sandra Susan. 2007 . Lone pursuit: Distrust and defensive individualism among the black poor . New York : Russell Sage Foundation . Stack C.B. 1974 . All our kin: Strategies for survival in a black community . New York : Basic Books . StataCorp . 2012 . “Stata Statistical Software: Release 13.” College Station, TX : Stata Corporation . Swaroop Sapna , Morenoff Jeffrey D. . 2006 . “Building community: The neighborhood context of social organization.” Social Forces 84 3 : 1665 - 95 . Van Hook J. , Balistreri K.S. . 2006 . “Ineligible parents, eligible children: Food Stamps receipt, allotments, and food insecurity among children of immigrants.” Social Science Research 35 1 : 228 - 51 . Vozoris Nicholas T. , Tarasuk Valerie S. . 2003 . “Household food insufficiency is associated with poorer health.” The Journal of Nutrition 133 1 : 120 -2 6 . Whitaker R. C. , Phillips S. M. , Orzol S. M. . 2006 . “Food insecurity and the risks of depression and anxiety in mothers and behavior problems in their preschool-aged children.” Pediatrics 118 3 : e859 - 68 . Wight Vanessa R. , Thampi Kalyani , Briggs Jodie . 2010 . “Who Are America's Poor Children? Examining Food Insecurity Among Children in the United States.” Columbia University, Mailman School of Public Health : National Center for Children in Poverty . Wilson William Julius. 1987 . The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy . Chicago, IL : University of Chicago Press . Wilson William Julius. 1996 . When Work Disappears: The World of the Urban Poor . New York : Alfred Knopf . Wodtke Geoffrey T. , Harding David J. , Elwert Relix . 2016 . “Neighborhood Effect Heterogeneity by Family Income and Developmental Period.” American Journal of Sociology 121 4 : 1168 - 222 . © The Author 2017. Published by Oxford University Press on behalf of the Society for the Study of Social Problems. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Social Problems Oxford University Press

Neighborhoods and Food Insecurity in Households with Young Children: A Disadvantage Paradox?

Social Problems , Volume 65 (3) – Aug 1, 2018

Loading next page...
 
/lp/ou_press/neighborhoods-and-food-insecurity-in-households-with-young-children-a-T5qtMcMHTP
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press on behalf of the Society for the Study of Social Problems. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
0037-7791
eISSN
1533-8533
D.O.I.
10.1093/socpro/spw054
Publisher site
See Article on Publisher Site

Abstract

Abstract In the United States, more than 1 in 5 households with children are unable to access and provide adequate food for a healthy, active lifestyle. We argue that the contribution of local context for food insecurity risk has largely been overlooked in favor of focusing on individual family characteristics, and that this is problematic given that mitigating food insecurity may be a communal process. We examine the relevance of neighborhood contributors to food insecurity among children, utilizing geocoded and nationally-representative data from the ECLS-K: 2010-2011 kindergarten cohort. We find little evidence that neighborhood socioeconomic, food retail, or social services characteristics directly impact food insecurity risk. However, our results reveal that family and neighborhood socioeconomic characteristics combine to impact food insecurity in ways consistent with a disadvantage paradox. As neighborhood concentrated disadvantage increases, higher-SES families’ risk of food insecurity increases, but lower-SES families’ risk decreases. This paradox is not explained by a higher concentration of social service organizations in more disadvantaged neighborhoods, and we theorize that impoverished families with children may share information and resources in disadvantaged communities to avoid food insecurity. food insecurity, SES, neighborhood disadvantage, ECLS-K, children Researchers, policymakers, and citizens alike agree that all children deserve the opportunity to develop and thrive. However, in the United States, one of the wealthiest nations in the world, 1 in 5 children live in poverty, including 24% of all children under the age of 6 (CLASP 2013). Poverty during childhood has immediate and long-term consequences for educational achievement, employment and earnings prospects, and health. Many children growing up in poverty experience these deficits as a result of living in a household that struggles with having enough food. Household food insecurity, a household’s collective inability to access adequate food for a healthy, active lifestyle, impacts nearly 16% of all households in the U. S. (Coleman-Jensen, Gregory and Singh 2014) and more than one in five households with children (Wight, Thampi and Briggs 2010). More troubling in this post-economic recession era, food insecurity is at the highest level of severity ever measured and has increased over 30% since 2007 despite federal food and nutrition assistance programs aimed at its elimination (Nord et al. 2010). Moreover, it is likely that estimates of the prevalence of food insecurity among households with children are underestimates – given that parents may be reluctant to admit their inability to provide enough food for their families, or are unaware that their children are suffering. One study of adolescents showed wide discrepancies between parent- and adolescent-reported accounts of food insecurity (Nord and Hanson 2012), and another mixed-methods study found a similar pattern for children’s accounts of food insecurity (Fram et al. 2013), such that adolescents and children report being food insecure at higher rates than do their parents. The precise nature of how to identify, examine, and alleviate food insecurity is complex. Strategies and current policies to address food insecurity among children have largely focused on families. While these policies may help individual families, to date they have struggled to curtail this enduring problem. We argue that one reason behind stalled progress may be failing to consider the community characteristics that might influence food insecurity, above and beyond individual or family factors. Existing knowledge of who and especially where food insecure children are remains quite limited, as well as how family and community characteristics might combine to influence the risk of food insecurity. We have three primary aims in this paper. First, we document how the neighborhoods of food secure and insecure children differ. Second, we investigate whether and how neighborhood characteristics influence the odds of food insecurity for these households by simultaneously considering the influence of family and neighborhood level characteristics. Finally, we assess how neighborhood characteristics might moderate the powerful influence of family-level SES on a child’s risk of living in a food insecure household. Food Insecurity Among Children Scholars interested in poverty and its effects on children should understand that household food insecurity is an indicator of material hardship that has far-reaching implications for children’s development and is social in nature. That is, it is a factor facing families which may be alleviated by collective and reciprocal assistance from entities outside the home: extended family members, other families, neighbors, or social and governmental organizations. Perhaps due in part to collective processes, poverty and food insecurity are strongly related but they are not synonymous. This is evidenced by the many poor families who remain food secure (Gundersen, Kreider, and Pepper 2011). Thus, for scholars interested in social processes related to disadvantage and how it might be mitigated, food insecurity is critical to understand. In this paper, we focus on food insecurity among households with children. Typically, the percentage of households with children who report food insecurity exceeds the percentage of food insecure children because in some households, parents will forego meals to ensure their children receive adequate food. Moreover, qualitative evidence demonstrates that even when parents believe they are shielding their children from the detrimental effects of food insecurity, children still suffer and in fact, take responsibility for implementing complex strategies of their own to mitigate food insecurity for their families (Fram et al. 2011). The consequences of food insecurity for children’s well-being are clearly established (Alaimo et al. 2001; Cook et al. 2004; Gundersen, Kreider and Pepper 2011). Existing research shows that the number of episodes of hunger that children experience is related to their health as they grow (Kirkpatrick, McIntyre and Potestio 2010), indicating that the healthy development of children is associated with not just the presence, but also the severity of food insecurity. Children who experience persistent food insecurity during early childhood, for instance, have worse health in adolescence relative to those who never or transiently experience food insecurity (Ryu and Bartfeld 2012). The costs of food insecurity in children extend beyond physical to mental health and academic performance. Detrimental performance in math and reading, loss of school days and repeated grades, behavior or attention problems, special education or mental health counseling, and suicidal or depressive tendencies among adolescents have all been linked to living in food insecure households (Alaimo, Olson and Frongillo 2002; Alaimo, Olson and Frongillo 2001; Jyoti, Frongillo and Jones 2005; Kimbro and Denney 2015; Kleinman et al. 1998; McIntyre et al. 2013; Murphy et al. 1998; Whitaker, Phillips and Orzol 2006). Much research has focused on the household-level characteristics associated with food insecurity. Family structure is an important predictor, with increased risk among households headed by single women (Nord et al. 2010). Maternal factors, including mental health and citizenship statuses (Van Hook and Balistreri 2006), also affect children’s risk of food insecurity. For example, children who have noncitizen mothers have levels of food insecurity nearly twice as high as those with native-born mothers (Kalil and Chen 2008), likely partially due to the families’ ineligibility (or belief of ineligibility) for federal food assistance. In addition, rates of food insecurity are higher than the national average among black and Hispanic households (Nord et al. 2010). As with a number of adult and child well-being outcomes, low socioeconomic status (SES), as indicated through a variety of measures, such as maternal education, family income, and participation in food assistance programs such as the Supplemental Nutrition Assistance Program (SNAP; formerly known as Food Stamps), WIC, or free/reduced breakfast and lunch, are strongly associated with household food insecurity, even after accounting for issues of selection (Alaimo et al. 1998; Casey et al. 2001; Jones et al. 2003; Kalil and Chen 2008; Rose and Richards 2004). However, at the same time, researchers have puzzled over how some disadvantaged families remain food secure (Nord et al. 2010). What strategies are poor families who remain food secure implementing? How are they shielding their children from harm? Below, we invoke interdisciplinary scholarship on poverty and discuss how some poor families may be relying on other external sources of resource allocation and support to avoid food insecurity, and argue that access to these external sources of assistance is not evenly distributed across communities. Neighborhoods and Well-being Though characteristics of children and their families motivate most research and policy strategies to curb food insecurity to date, there is ample evidence that neighborhood factors influence child well-being (Leventhal and Brooks-Gunn 2000). We conceptualize communities as influencing the prospects for healthy development because of the institutional and social resources immediately available (or unavailable) to residents (Jencks and Mayer 1990). Neighborhoods with more wealth tend to attract more institutional resources such as retail establishments, parks, and other amenities. Higher-SES neighborhoods also may be characterized by a greater density of other institutional resources like schools, libraries, and community organizations, the lack of which may contribute to the impact of neighborhood disadvantage on residents’ well-being (Wilson 1996). The social resources available in a neighborhood, or community social organization, may also impact food insecurity in a variety of ways (Crowe and Smith 2012). On the one hand, more advantaged areas may have higher levels of social integration, provide more secure and beneficial social networks, an enhanced sense of social capital for residents, and consequently represent places where children are more likely to thrive (Sampson 2012). Neighborhoods rich in social ties and with norms of reciprocity are likely to be places where neighbors look out for each other. In contrast, neighborhoods where distrust pervades the community are places where social capital has difficulty being generated or sustained, and neighbors may keep to themselves (Rankin and Quane 2000). Social resources can take at least two forms in communities – ties to other individuals or families, and ties to social organizations. In more disadvantaged neighborhoods, residents are less likely to have employed friends and more likely to have friends on public assistance (Rankin and Quane 2000), and better-off residents may minimize their associations with lower-income neighbors (Anderson 1990). Ties to civic institutions and social organizations may also be diminished in some (Wilson 1987), though importantly not all (Klinenberg 2002; Small 2004), lower-income neighborhoods. In all, social and economic disadvantages may extend beyond individual or family disadvantages to area-level disadvantages to influence well-being. On the other hand, linking neighborhood socioeconomic disadvantage to individual well-being is more complicated. Though there are several examples to suggest that poor residents in poor neighborhoods face special heightened threats to well-being (Wilson 1987; Wilson 1996), scholars have shown time and again that not all poor communities are alike (Klinenberg 2002; Seccombe 2002; Sherman 2006; Small 2004; Small, Harding and Lamont 2010). In spite of depictions of poor communities as devoid of social capital with residents destined for poor health and well-being, there are clear exceptions that reveal strategies to persevere in the face of seemingly overwhelming adversity (Desmond 2012; Klinenberg 2002; Seccombe 2002; Small 2004; Stack 1974). Accordingly, there is value in assessing whether well-being differs for more and less disadvantaged families who reside in neighborhoods of different levels of socioeconomic status. Neighborhoods and Food Insecurity To date, social scientists have paid little attention to the relationship between social and economic neighborhood environments, including the availability of food and social service agencies, and food insecurity. As a notable exception, Kirkpatrick and Tarasuk’s (2010; 2011) research evaluates the importance of both food environments and perceived neighborhood social capital on families’ food insecurity risks in an urban area in Canada. The authors assess the distance to food outlets and neighborhood social capital and conclude that the causes of food insecurity lie with characteristics of households and not necessarily neighborhoods. These results stand in contrast to neighborhood effects literatures (see Kawachi and Berkman 2003) which suggest that area-level characteristics shape individual well-being. A recent descriptive U.S. study suggests that structural factors at the neighborhood level, such as aggregate levels of poverty, may impact food insecurity above and beyond household-level factors (Kimbro, Denney and Panchang 2012). Neighborhood disadvantage as it may relate to household food insecurity can be described in multiple ways. First, social and economic characteristics such as the neighborhood proportion in poverty, proportion unemployed, proportion of female-headed households, and the proportion receiving public assistance collectively represent variation in what has been termed concentrated disadvantage across neighborhoods (Morenoff 2003; Sampson, Raudenbush and Earls 1997; Swaroop and Morenoff 2006). Second, food availability in neighborhoods can be measured more directly by assessing the neighborhood availability of supermarkets and convenience stores. One assumption is that a greater availability of grocery stores could reduce food insecurity, particularly in low-income neighborhoods with few or no supermarkets. Conversely, a high representation of convenience stores might indicate a lack of quality foods. The literature on the link between grocery store access and food security is international in scope and mixed in findings, with some studies reporting positive associations and others reporting no association (see Sadler, Gilliland and Arku 2013 for a discussion). Recent work in Flint, Michigan finds no significant association between physical access to food and household food security (Sadler, Gilliland and Arku 2013). Similarly, another recent U.S. study in Philadelphia found no association between grocery store availability and obesity (Cummins, Flint and Matthews 2014). The recent evidence suggests that food access environments may not be as influential on populations as once perceived (Sadler, Gilliland and Arku 2013). However, these recent studies focused on single communities, so it remains unclear if food access nationwide is a significant predictor of household food insecurity. Finally, measurement of social services such as the availability of food pantries within neighborhoods may associate with household food insecurity over and above family level characteristics. One possibility is that social services aimed at alleviating food insecurity may be more prevalent in disadvantaged neighborhoods featuring residents most in need of assistance. A first set of hypotheses investigates the relationship between neighborhood indicators of disadvantage and household food insecurity over and above the relationships between family status and food insecurity: H1: Neighborhood concentrated disadvantage will be associated with higher odds of household food insecurity. H2: Greater supermarket availability in neighborhoods will be associated with lower, and higher proportions of convenience stores will be associated with higher, odds of household food insecurity. H3: Greater representation of social assistance organizations will be associated with lower odds of household food insecurity. These hypotheses assess if characteristics of the communities where children live influence the odds of food insecurity for the household but they do not specify how, and for whom, neighborhoods might matter. Below, we describe at least two possibilities aimed at understanding how neighborhoods might matter specifically for children in the most disadvantaged families: 1) accumulating household and neighborhood disadvantages and 2) a disadvantage paradox, whereby the most disadvantaged households may experience lower risks of food insecurity the more disadvantaged the neighborhood. Accumulating Risks Decades of neighborhood research on important social problems such as health (Kawachi and Berkman 2003), safety (Sampson, Raudenbush and Earls 1997), and educational attainment (Borman and Dowling 2010; Jencks and Mayer 1990), stipulate that disadvantages across individuals and the places they live possess the capability to accumulate. For example, a recent and sophisticated study demonstrated that living in a disadvantaged neighborhood is most consequential for high school graduation for poor children (Wodtke, Harding and Elwert 2016). In the context of food insecurity, an accumulation of risk might include places with few jobs, little transportation, and/or few supermarkets that provide healthy foods at lower costs and which accept assistance programs such as SNAP. These same places disproportionately include individuals and families that possess fewer resources who would rely more heavily on neighborhood resources were they available. Thus, the contributions of household and neighborhood level disadvantages might synergistically influence risk. If so, then neighborhood disadvantages increase the risk of food insecurity more for more socioeconomically disadvantaged households than for less disadvantaged households. Thus, H4: The risk of food insecurity for children in the most disadvantaged families will increase the more disadvantaged the neighborhood. A Disadvantage Paradox Conceptualizing food insecurity, especially for children, as an outcome with collective or community implications reveals the possibility of a unique interaction between neighborhoods and families. Not all poor communities are alike (Crowe and Smith 2012; Klinenberg 2002; Small, Harding and Lamont 2010). Structurally, some poor neighborhoods possess resources such as community centers or food pantries specifically aimed at alleviating, and in response to, disadvantage. These community resources provide the potential to help all residents, regardless of individual SES (Carpiano, Lloyd and Hertzman 2009), and some evidence suggests that participation in social organizations may be higher in more disadvantaged neighborhoods than in less disadvantaged ones (Rankin and Quane 2000). Further, neighborhood characteristics comprise indicators of social networks and connectivity that come with links to resources or information (Carpiano 2008; Kawachi 2010; Small 2009; Swaroop and Morenoff 2006). Given that healthy foods are not equally accessible across all communities (Hung 1999; Jetter and Cassady 2006; Morland, Diez Roux and Wing 2006), community resources and support that can be leveraged to avoid food insecurity may also fluctuate. Qualitative research concludes that insufficient food supplies are likely not the result of some deficiency at the household level but rather extend to wider social networks and the social and economic characteristics of the communities in which families live (Ahluwalia, Dodds and Baligh 1998). There is also evidence to suggest that greater civic structure within communities can reduce the chance of food insecurity (De Marco and Thorburn 2009; Morton et al. 2005; Vozoris and Tarasuk 2003). A small study in Connecticut found that social capital at the household and community level was associated with a reduced risk of food insecurity (Martin et al. 2004). Indeed, researchers have documented that disadvantaged persons and families pool resources with each other to endure difficult circumstances (Morton et al. 2008; Stack 1974). Individuals in poor communities will respond to deprivation in varying ways as well, influenced in part by cultural expectations and moral rules informally and formally enacted by area residents (Anderson 1999; Sherman 2006). Coping strategies in poor neighborhoods that are viewed as more in line with expectations of behavior may generate what Jennifer Sherman (2006) refers to as “moral capital.” In turn, moral capital can be traded for “…social capital in the form of community ties or social support” (Sherman 2006: 893). Though Sherman identifies this form of social resource sharing in rural communities dealing with poverty, Lichter and Brown (2011) provide reasons to believe that processes occurring in rural spaces might under some circumstances also occur in urban settings as the grip of poverty does not discriminate spatially, nor do collective reactions to it. Recent research on the working poor and the Earned Income Tax Credit (EITC) bears this out. Indeed, poor residents almost uniformly reject the idea that the EITC is a form of government assistance but rather, as the name implies, earned income (Halpern-Meekin et al. 2015). Thus, applying for and receiving the EITC to help with ongoing economic struggles is not only a survival strategy but perhaps the moral and right thing to do, creating a cultural space where bonds are formed and values shared. In addition, disadvantaged groups at times exhibit more prosocial resource allocation and a greater tendency toward egalitarian social values than more advantaged persons and families (Piff et al. 2010). In the case of households with young children, these phenomena may be particularly pronounced. Establishing strong networks can be seen as an adaptive strategy used to deal with the consequences of poverty (Dominguez and Watkins 2003; Menjivar 2000; Newman 1992). Although many scholars believe the kin networks Stack (1974) observed have all but disappeared in the modern disadvantaged neighborhood (see Smith 2007), Desmond (2012) writes that they have been replaced with quickly-formed and quickly-dissolved intense ties which enable survival. Finally, evidence suggests that if these processes occur in disadvantaged neighborhoods they are likely to apply most to the most disadvantaged residents of the area. Anderson (1990) witnessed patterns of behavior in disadvantaged neighborhoods where more advantaged residents minimized their associations with lower-income neighbors. It may be that better-off residents do not agree with the coping strategies employed by their struggling counterparts and thus remove themselves from interactions that might generate moral and social capital (Sherman 2006). This stream of research provides an alternative to the accumulation of disadvantage hypothesis: H5: The risk of food insecurity for children in the most disadvantaged families will decrease the more disadvantaged the neighborhood. By comparison, for both the accumulation of risks and the disadvantage paradox perspectives, children in less disadvantaged families may experience higher risks of food insecurity the more disadvantaged the neighborhood. But if the accumulation perspective is supported, then those increases will be less pronounced for children in less disadvantaged families compared to children in the most disadvantaged families. Alternatively, if the paradox is supported, then we should see a reduction in risk as neighborhood disadvantage increases for more disadvantaged children; and an increase in risk for less disadvantaged children. DATA AND METHOD This study uses restricted, geo-coded data from the spring kindergarten wave of the Early Childhood Longitudinal Study-Kindergarten Class of 2010-2011 (ECLS-K), which is a nationally-representative sample of U.S. children who were in kindergarten in 2010-2011. Food insecurity was collected only in the spring data collection, so we utilize data primarily from the spring kindergarten wave. The restricted version of ECLS-K provides census tract information with which data from the American Community Survey (2007-2011) are appended to create neighborhood-level measures. We utilize the ACS instead of decennial Census 2010 now that the ACS routinely collects long-form information on socioeconomic characteristics such as poverty status. We also link the ECLS-K to 2010 Zip Code Business Patterns data to create estimates of the density of food retail and social service organizations in census tracts.1 The ECLS-K sample includes 18,100 children (in accordance with our restricted data agreement we round all sample sizes to the nearest 50) with valid geocodes at the spring kindergarten wave. About 29% of this sample is missing a parent interview at the spring wave, when food insecurity was collected, so our final sample size is 12,800 children who live in 3,800 Census tracts. Roughly 10% of the sample represents the sole observation in their census tract and the average number of children per census tract is 3.4; we find no evidence this creates estimation problems (Bell, Ferron and Kromrey 2008). But we do conduct sensitivity analyses by comparing our results before and after dropping the singleton neighborhood cases and find no differences in results so they are included in the final models. Variables Our outcome measure is a dichotomous measure of household food insecurity derived from the USDA’s 18-item food insecurity scale (Bickel et al. 2000). Examples of the questions include, “In the last 12 months, were you ever hungry but didn’t eat because you couldn’t afford enough food?” and “In the last 12 months, was [child] ever hungry but you just couldn’t afford more food?” Following USDA guidelines (2000), we classify the households of parents answering in the affirmative to three or more of the items as food insecure. Individual-level variables include the child’s age in months, race/ethnicity (non-Hispanic white, non-Hispanic black, non-Hispanic other, and Hispanic), and gender (1 = male); the mother’s age and nativity status (1 = foreign born); the number of siblings in the household and family structure (two-parent family, single-mother family, and other family type). We incorporate a series of health and well-being measures, which are asked of the primary caregiver of the child (the biological mother in most cases). The ECLS-K utilizes an abbreviated version of the 20-item Center for Epidemiological Studies of Depression Scale (CES-D) (Radloff 1977). Following convention, we classify mothers as “likely depressed” if her total score on the abbreviated scale is equal to or greater than 12. We include dichotomous indicators for whether the child and primary caregiver are in “fair or poor health,” utilizing the standard five-response self-rated health measure, as well as an indicator for whether the primary caregiver reports health-related work limitations. We also account for whether the household currently receives SNAP benefits, as well as the length of time the child has lived in the current residence (measured in years) – note that this measure was collected only in the fall kindergarten wave. To measure family SES, we utilize the composite measure provided by the ECLS, which incorporates household income, both parents’ occupational prestige, and both parents’ education (this measure incorporates the custodial parent’s data if only one parent is present in the household). Each component is standardized, and the resulting z-scores are added and divided by the number of components to yield an overall standardized measure of SES. To mirror interpretation for our neighborhood disadvantage measure described below, we reverse-coded this measure so that higher values indicate greater socioeconomic adversity. In other words, a value of 1 indicates a family is one standard deviation higher in socioeconomic adversity than the mean. To capture the neighborhood characteristics that may influence food insecurity risk we control for neighborhood population (logged) and a dichotomous indicator for whether the child lives in an urban area. For social and economic disadvantage in neighborhoods we include an index of concentrated disadvantage, created based on the first dimension of a principal components factor analysis on percent of adults in the census tract living below the poverty line, the percent of households receiving public assistance, the percent of adult residents who are unemployed, and the percent of female-headed households with children (Sampson, Raudenbush and Earls 1997). We also include two measures of the food retail environment in a neighborhood, the density of large supermarkets (50+ employees) and convenience stores per square mile. Finally, we include a measure of the density of social service organizations in the neighborhood, all measured per square mile – social assistance organizations related to children and youth (e.g. Boys and Girls Clubs, AFDC, day care), community services (e.g. housing shelters, food pantries), and vocational rehabilitation services (e.g., job training and counseling). These measures are also logged in our models. Missing Data Approximately 20% of children remaining in our analytic sample are missing data on one or more measures of interest, with the vast majority (80%) of those missing only on the number of years the child had lived at the current residence (because it was only collected in the fall kindergarten wave, so children missing a parent interview in the fall are missing on this item). Children missing data were more often non-Hispanic black or Hispanic, poorer, lived with single mothers, and were more likely to have foreign-born parents. Given the evidence that our missing data are not missing completely at random and may be conditioned by other observed covariates, standard procedures such as listwise deletion would be inappropriate (Allison 2002). Instead, we use multiple imputation procedures in Stata 12 (Royston 2005) to estimate values for our multivariate analyses. During imputation, a diverse set of predictors estimate five sets of probable values for each missing value. The resulting five data sets include a random component based on draws from the posterior predictive distribution of the missing data under a posited Bayesian model and provide unbiased estimates of variance (Allison 2002). Models estimated without imputation provide results very similar to those using imputation (available upon request). Estimation To test the effects of neighborhood conditions on odds of food insecurity among households with children, we estimate multi-level logistic regression models (Guo and Zhao 2000; Rabe-Hesketh and Skrondal 2008) using the mi estimate command within Stata 12 software (StataCorp 2012). The models treat level-1 children as nested within level-2 census tracts. All models utilize maximum likelihood estimation with adaptive quadrature (Rabe-Hesketh and Skrondal 2008), adjusting for clustering by neighborhood, different sample sizes for level-1 and level-2 units, heteroscedastic error terms, and varying numbers of cases within level-2 units – all problems that otherwise downwardly bias estimated standard errors (Raudenbush and Bryk 2002). We evaluate hypotheses 1 through 3 by including neighborhood-level (level-2) predictors one at a time along with the level-1 predictors. Hypotheses 4 and 5 specify cross-level interactions between family SES at level-1 and concentrated disadvantage, food retail, and social organizations at level-2. The error terms associated with the interaction models are assumed to be multivariate normally distributed, each with a mean of 0 and non-zero variances and covariances. We report all regression results as odds ratios (OR) and generate predicted probabilities to illustrate findings from the cross-level interactions. RESULTS First, Table 1 provides weighted means and proportions for the dependent and independent variables at both the individual/family level and at the neighborhood level. In addition, the table provides significance tests of differences for households reporting food insecurity versus those not reporting food insecurity. Table 1. Means and Proportions, ECLS-K: 2010-2011 Data Panel A. Individual and Family Characteristics n = 12,800 Full Sample Food Secure Food Insecurea Food insecure 0.12 Child is male 0.51 0.51 0.53 Child's age in months 74.60 74.50 74.80* Foreign-born parent 0.34 0.33 0.44*** Race/ethnicity  non-Hispanic white (ref) 0.51 0.54 0.34***  non-Hispanic black 0.11 0.10 0.15***  Hispanic 0.24 0.21 0.40***  Other 0.14 0.15 0.11*** Mother's age in years 34.50 34.70 33.20*** Family Structure  Two biological parents, married or cohab. (ref) 0.72 0.74 0.57***  Single mother family 0.20 0.18 0.33***  Other family type 0.08 0.08 0.10+ Family socioeconomic adversity (standardized) 0.05 −0.09 0.61*** Family receives food stamps 0.26 0.22 0.54*** Number of siblings in the household 1.46 1.42 1.75*** Length of time child has lived in current house (years) 2.77 2.81 2.49*** Child in fair/poor health 0.02 0.02 0.06*** Primary caregiver likely depressed 0.08 0.06 0.24*** Primary caregiver in fair/poor health 0.10 0.08 0.25*** Primary caregiver has health-related work limitations 0.07 0.06 0.16*** Panel B. Neighborhood Characteristics n = 3,800 Full Not Food Insecure Food Insecurea Neighborhood is urban 0.77 0.76 0.82*** Concentrated Disadvantage (standardized) 0.00 −0.13 0.33***  (components of concentrated disadvantage)   % Households in poverty 0.13 0.12 0.16***   % Female headed household 0.14 0.13 0.16***   % unemployed 0.09 0.08 0.10***   % Residents receiving public assistance 0.03 0.02 0.03*** Neighborhood Resource Measures (density per square mile)  Large supermarkets (50+ employees) 0.28 0.26 0.29*  Convenience Stores 0.46 0.39 0.53***  Social assistance organizations 2.89 2.47 3.13** Panel A. Individual and Family Characteristics n = 12,800 Full Sample Food Secure Food Insecurea Food insecure 0.12 Child is male 0.51 0.51 0.53 Child's age in months 74.60 74.50 74.80* Foreign-born parent 0.34 0.33 0.44*** Race/ethnicity  non-Hispanic white (ref) 0.51 0.54 0.34***  non-Hispanic black 0.11 0.10 0.15***  Hispanic 0.24 0.21 0.40***  Other 0.14 0.15 0.11*** Mother's age in years 34.50 34.70 33.20*** Family Structure  Two biological parents, married or cohab. (ref) 0.72 0.74 0.57***  Single mother family 0.20 0.18 0.33***  Other family type 0.08 0.08 0.10+ Family socioeconomic adversity (standardized) 0.05 −0.09 0.61*** Family receives food stamps 0.26 0.22 0.54*** Number of siblings in the household 1.46 1.42 1.75*** Length of time child has lived in current house (years) 2.77 2.81 2.49*** Child in fair/poor health 0.02 0.02 0.06*** Primary caregiver likely depressed 0.08 0.06 0.24*** Primary caregiver in fair/poor health 0.10 0.08 0.25*** Primary caregiver has health-related work limitations 0.07 0.06 0.16*** Panel B. Neighborhood Characteristics n = 3,800 Full Not Food Insecure Food Insecurea Neighborhood is urban 0.77 0.76 0.82*** Concentrated Disadvantage (standardized) 0.00 −0.13 0.33***  (components of concentrated disadvantage)   % Households in poverty 0.13 0.12 0.16***   % Female headed household 0.14 0.13 0.16***   % unemployed 0.09 0.08 0.10***   % Residents receiving public assistance 0.03 0.02 0.03*** Neighborhood Resource Measures (density per square mile)  Large supermarkets (50+ employees) 0.28 0.26 0.29*  Convenience Stores 0.46 0.39 0.53***  Social assistance organizations 2.89 2.47 3.13** Source: ECLS-K: 2010-2011 (N = 12,800 in 3,800 neighborhoods); 2007-2010 ACS data at the tract level a p-value for a t-test of significance by column (i.e. not food insecure vs. food insecure); *** ≤ 0.001 **≤ 0.01 *≤ 0.05 + ≤ 0.10 Table 1. Means and Proportions, ECLS-K: 2010-2011 Data Panel A. Individual and Family Characteristics n = 12,800 Full Sample Food Secure Food Insecurea Food insecure 0.12 Child is male 0.51 0.51 0.53 Child's age in months 74.60 74.50 74.80* Foreign-born parent 0.34 0.33 0.44*** Race/ethnicity  non-Hispanic white (ref) 0.51 0.54 0.34***  non-Hispanic black 0.11 0.10 0.15***  Hispanic 0.24 0.21 0.40***  Other 0.14 0.15 0.11*** Mother's age in years 34.50 34.70 33.20*** Family Structure  Two biological parents, married or cohab. (ref) 0.72 0.74 0.57***  Single mother family 0.20 0.18 0.33***  Other family type 0.08 0.08 0.10+ Family socioeconomic adversity (standardized) 0.05 −0.09 0.61*** Family receives food stamps 0.26 0.22 0.54*** Number of siblings in the household 1.46 1.42 1.75*** Length of time child has lived in current house (years) 2.77 2.81 2.49*** Child in fair/poor health 0.02 0.02 0.06*** Primary caregiver likely depressed 0.08 0.06 0.24*** Primary caregiver in fair/poor health 0.10 0.08 0.25*** Primary caregiver has health-related work limitations 0.07 0.06 0.16*** Panel B. Neighborhood Characteristics n = 3,800 Full Not Food Insecure Food Insecurea Neighborhood is urban 0.77 0.76 0.82*** Concentrated Disadvantage (standardized) 0.00 −0.13 0.33***  (components of concentrated disadvantage)   % Households in poverty 0.13 0.12 0.16***   % Female headed household 0.14 0.13 0.16***   % unemployed 0.09 0.08 0.10***   % Residents receiving public assistance 0.03 0.02 0.03*** Neighborhood Resource Measures (density per square mile)  Large supermarkets (50+ employees) 0.28 0.26 0.29*  Convenience Stores 0.46 0.39 0.53***  Social assistance organizations 2.89 2.47 3.13** Panel A. Individual and Family Characteristics n = 12,800 Full Sample Food Secure Food Insecurea Food insecure 0.12 Child is male 0.51 0.51 0.53 Child's age in months 74.60 74.50 74.80* Foreign-born parent 0.34 0.33 0.44*** Race/ethnicity  non-Hispanic white (ref) 0.51 0.54 0.34***  non-Hispanic black 0.11 0.10 0.15***  Hispanic 0.24 0.21 0.40***  Other 0.14 0.15 0.11*** Mother's age in years 34.50 34.70 33.20*** Family Structure  Two biological parents, married or cohab. (ref) 0.72 0.74 0.57***  Single mother family 0.20 0.18 0.33***  Other family type 0.08 0.08 0.10+ Family socioeconomic adversity (standardized) 0.05 −0.09 0.61*** Family receives food stamps 0.26 0.22 0.54*** Number of siblings in the household 1.46 1.42 1.75*** Length of time child has lived in current house (years) 2.77 2.81 2.49*** Child in fair/poor health 0.02 0.02 0.06*** Primary caregiver likely depressed 0.08 0.06 0.24*** Primary caregiver in fair/poor health 0.10 0.08 0.25*** Primary caregiver has health-related work limitations 0.07 0.06 0.16*** Panel B. Neighborhood Characteristics n = 3,800 Full Not Food Insecure Food Insecurea Neighborhood is urban 0.77 0.76 0.82*** Concentrated Disadvantage (standardized) 0.00 −0.13 0.33***  (components of concentrated disadvantage)   % Households in poverty 0.13 0.12 0.16***   % Female headed household 0.14 0.13 0.16***   % unemployed 0.09 0.08 0.10***   % Residents receiving public assistance 0.03 0.02 0.03*** Neighborhood Resource Measures (density per square mile)  Large supermarkets (50+ employees) 0.28 0.26 0.29*  Convenience Stores 0.46 0.39 0.53***  Social assistance organizations 2.89 2.47 3.13** Source: ECLS-K: 2010-2011 (N = 12,800 in 3,800 neighborhoods); 2007-2010 ACS data at the tract level a p-value for a t-test of significance by column (i.e. not food insecure vs. food insecure); *** ≤ 0.001 **≤ 0.01 *≤ 0.05 + ≤ 0.10 Approximately 12% of children in the ECLS-K Spring sample lived in households that met the criteria for food insecurity. Food insecurity in households with children is statistically similar for male and female children. Otherwise, important descriptive differences exist at the family level. Compared to children in food secure households, children in food insecure households are older, more likely to have a foreign-born parent, more likely black or Hispanic, have lived a shorter time at their current address, and have more siblings. More dramatic differences exist in the health status of the caregiver, in family structure, and in the socioeconomic situation of households. Roughly 6 and 8% of caregivers in food secure households are likely depressed and in fair or poor health, respectively. In food insecure households, a full 25% of caregivers are likely depressed and/or in fair or poor health. Further, about 18% of food secure households are single mother families whereas a third of all food insecure households are single mother families. Finally, food insecure households are, as expected, much more disadvantaged socioeconomically than are food secure households. For example, over half of food insecure households receive food stamps whereas less than a quarter of food secure households receive food stamps. Table 1 also shows that there are significant differences in the neighborhood characteristics of food insecure and food secure households. Children in food insecure households are more likely to live in more socially and economically disadvantaged neighborhoods. For example, children in food insecure households live in neighborhoods where on average 16% of all households are in poverty, compared to 12% for children in food secure households. Surprisingly, but consistent with other recent research (Sadler, Gilliland and Arku 2013), they are also located in neighborhoods with more supermarkets and with higher densities of convenience stores. This is partly a reflection of a larger proportion of ECLS-K children in urban than in rural neighborhoods overall. Indeed, 77% of children in the ECLS-K live in an urban neighborhood. Children in food insecure households are also more likely to live in neighborhoods with more social assistance organizations. To examine the relevance of neighborhood contributors for household food insecurity, we turn to the regression results in Table 2. Table 2. Multilevel Logistic Regression Models for Household Food Insecurity: Associations with Neighborhood Characteristics (Odds Ratios) Model 1 Model 2 Model 3 Model 4 Model 5 OR OR OR OR OR Child and Family Characteristics Child is male 1.04 1.04 1.04 1.04 1.04 Child's age in months 1.01 1.01 1.01 1.01 1.01 Foreign-born parent 1.26** 1.26** 1.27** 1.25** 1.26** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.86 0.85 0.86 0.85 0.86  Hispanic 1.06 1.05 1.07 1.06 1.06  Other 0.91 0.91 0.92 0.91 0.92 Mother's age in years 1.01* 1.01* 1.01* 1.01* 1.01* Family socioeconomic adversity (SEA) 2.33*** 2.32*** 2.33*** 2.33*** 2.33*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.14+ 1.15+ 1.15+ 1.15+  Other family type 0.82+ 0.82+ 0.82+ 0.82+ 0.82+ Number of siblings in the household 1.18*** 1.17*** 1.17*** 1.18*** 1.18*** Years lived in home 0.96* 0.96* 0.96* 0.96* 0.96* Family receives food stamps 1.66*** 1.66*** 1.66*** 1.66*** 1.66*** Child in fair/poor health 1.58** 1.58** 1.58** 1.58** 1.58** Primary caregiver likely depressed 2.89*** 2.89*** 2.89*** 2.89*** 2.89*** Primary caregiver in fair/poor health 1.58*** 1.57*** 1.58*** 1.58*** 1.58*** Primary caregiver has work limitations 1.54*** 1.54*** 1.54*** 1.54*** 1.54*** Neighborhood Characteristics Total Population (ln) 0.87* 0.86* 0.87+ 0.87* 0.87* Urban area 1.27** 1.27** 1.39*** 1.24* 1.31* Concentrated disadvantage 1.01 Neighborhood resources (density per square mile)  Large supermarkets (50+ employees) 0.97+  Convenience stores 1.01  Social assistance organizations 0.99 Constant 0.06*** 0.07*** 0.05*** 0.07*** 0.06*** sd(random intercept) 0.32*** 0.33*** 0.33*** 0.32*** 0.32*** sd(random slope - Family SEA) 0.14** 0.14** 0.14** 0.14** 0.14** Model 1 Model 2 Model 3 Model 4 Model 5 OR OR OR OR OR Child and Family Characteristics Child is male 1.04 1.04 1.04 1.04 1.04 Child's age in months 1.01 1.01 1.01 1.01 1.01 Foreign-born parent 1.26** 1.26** 1.27** 1.25** 1.26** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.86 0.85 0.86 0.85 0.86  Hispanic 1.06 1.05 1.07 1.06 1.06  Other 0.91 0.91 0.92 0.91 0.92 Mother's age in years 1.01* 1.01* 1.01* 1.01* 1.01* Family socioeconomic adversity (SEA) 2.33*** 2.32*** 2.33*** 2.33*** 2.33*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.14+ 1.15+ 1.15+ 1.15+  Other family type 0.82+ 0.82+ 0.82+ 0.82+ 0.82+ Number of siblings in the household 1.18*** 1.17*** 1.17*** 1.18*** 1.18*** Years lived in home 0.96* 0.96* 0.96* 0.96* 0.96* Family receives food stamps 1.66*** 1.66*** 1.66*** 1.66*** 1.66*** Child in fair/poor health 1.58** 1.58** 1.58** 1.58** 1.58** Primary caregiver likely depressed 2.89*** 2.89*** 2.89*** 2.89*** 2.89*** Primary caregiver in fair/poor health 1.58*** 1.57*** 1.58*** 1.58*** 1.58*** Primary caregiver has work limitations 1.54*** 1.54*** 1.54*** 1.54*** 1.54*** Neighborhood Characteristics Total Population (ln) 0.87* 0.86* 0.87+ 0.87* 0.87* Urban area 1.27** 1.27** 1.39*** 1.24* 1.31* Concentrated disadvantage 1.01 Neighborhood resources (density per square mile)  Large supermarkets (50+ employees) 0.97+  Convenience stores 1.01  Social assistance organizations 0.99 Constant 0.06*** 0.07*** 0.05*** 0.07*** 0.06*** sd(random intercept) 0.32*** 0.33*** 0.33*** 0.32*** 0.32*** sd(random slope - Family SEA) 0.14** 0.14** 0.14** 0.14** 0.14** *** < 0.001 ** < 0.01 * < 0.05 + < 0.10 Table 2. Multilevel Logistic Regression Models for Household Food Insecurity: Associations with Neighborhood Characteristics (Odds Ratios) Model 1 Model 2 Model 3 Model 4 Model 5 OR OR OR OR OR Child and Family Characteristics Child is male 1.04 1.04 1.04 1.04 1.04 Child's age in months 1.01 1.01 1.01 1.01 1.01 Foreign-born parent 1.26** 1.26** 1.27** 1.25** 1.26** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.86 0.85 0.86 0.85 0.86  Hispanic 1.06 1.05 1.07 1.06 1.06  Other 0.91 0.91 0.92 0.91 0.92 Mother's age in years 1.01* 1.01* 1.01* 1.01* 1.01* Family socioeconomic adversity (SEA) 2.33*** 2.32*** 2.33*** 2.33*** 2.33*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.14+ 1.15+ 1.15+ 1.15+  Other family type 0.82+ 0.82+ 0.82+ 0.82+ 0.82+ Number of siblings in the household 1.18*** 1.17*** 1.17*** 1.18*** 1.18*** Years lived in home 0.96* 0.96* 0.96* 0.96* 0.96* Family receives food stamps 1.66*** 1.66*** 1.66*** 1.66*** 1.66*** Child in fair/poor health 1.58** 1.58** 1.58** 1.58** 1.58** Primary caregiver likely depressed 2.89*** 2.89*** 2.89*** 2.89*** 2.89*** Primary caregiver in fair/poor health 1.58*** 1.57*** 1.58*** 1.58*** 1.58*** Primary caregiver has work limitations 1.54*** 1.54*** 1.54*** 1.54*** 1.54*** Neighborhood Characteristics Total Population (ln) 0.87* 0.86* 0.87+ 0.87* 0.87* Urban area 1.27** 1.27** 1.39*** 1.24* 1.31* Concentrated disadvantage 1.01 Neighborhood resources (density per square mile)  Large supermarkets (50+ employees) 0.97+  Convenience stores 1.01  Social assistance organizations 0.99 Constant 0.06*** 0.07*** 0.05*** 0.07*** 0.06*** sd(random intercept) 0.32*** 0.33*** 0.33*** 0.32*** 0.32*** sd(random slope - Family SEA) 0.14** 0.14** 0.14** 0.14** 0.14** Model 1 Model 2 Model 3 Model 4 Model 5 OR OR OR OR OR Child and Family Characteristics Child is male 1.04 1.04 1.04 1.04 1.04 Child's age in months 1.01 1.01 1.01 1.01 1.01 Foreign-born parent 1.26** 1.26** 1.27** 1.25** 1.26** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.86 0.85 0.86 0.85 0.86  Hispanic 1.06 1.05 1.07 1.06 1.06  Other 0.91 0.91 0.92 0.91 0.92 Mother's age in years 1.01* 1.01* 1.01* 1.01* 1.01* Family socioeconomic adversity (SEA) 2.33*** 2.32*** 2.33*** 2.33*** 2.33*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.14+ 1.15+ 1.15+ 1.15+  Other family type 0.82+ 0.82+ 0.82+ 0.82+ 0.82+ Number of siblings in the household 1.18*** 1.17*** 1.17*** 1.18*** 1.18*** Years lived in home 0.96* 0.96* 0.96* 0.96* 0.96* Family receives food stamps 1.66*** 1.66*** 1.66*** 1.66*** 1.66*** Child in fair/poor health 1.58** 1.58** 1.58** 1.58** 1.58** Primary caregiver likely depressed 2.89*** 2.89*** 2.89*** 2.89*** 2.89*** Primary caregiver in fair/poor health 1.58*** 1.57*** 1.58*** 1.58*** 1.58*** Primary caregiver has work limitations 1.54*** 1.54*** 1.54*** 1.54*** 1.54*** Neighborhood Characteristics Total Population (ln) 0.87* 0.86* 0.87+ 0.87* 0.87* Urban area 1.27** 1.27** 1.39*** 1.24* 1.31* Concentrated disadvantage 1.01 Neighborhood resources (density per square mile)  Large supermarkets (50+ employees) 0.97+  Convenience stores 1.01  Social assistance organizations 0.99 Constant 0.06*** 0.07*** 0.05*** 0.07*** 0.06*** sd(random intercept) 0.32*** 0.33*** 0.33*** 0.32*** 0.32*** sd(random slope - Family SEA) 0.14** 0.14** 0.14** 0.14** 0.14** *** < 0.001 ** < 0.01 * < 0.05 + < 0.10 The results for Model 1 of Table 2 are largely in line with the body of empirical research on food insecurity, showing that characteristics of children and their families are strongly associated with the odds of household food insecurity. That is, children who are older, have a foreign-born parent (relative to those without), have older mothers, and have primary caregivers who report health and disability issues have higher odds of food insecurity. In particular, if the caregiver is likely depressed, the household in which the child lives has nearly three times higher odds of food insecurity, relative to a household where the caregiver is not likely depressed. Finally, families at greater socioeconomic adversity face significantly higher odds of food insecurity. Controlling for all other child and family characteristics, a one standard deviation increase in socioeconomic adversity is associated with 2.3 times higher odds of household food insecurity. Models 2-5 show that these child and family associations with household food insecurity are robust and go largely unchanged as neighborhood characteristics enter the model. More important for our purposes, we find no significant associations with food insecurity and neighborhood concentrated disadvantage (Model 2), density of supermarkets or convenience stores (Models 3 and 4), or social assistance organizations (Model 5). Thus, we find no support for hypotheses 1 to 3 – neighborhood characteristics are not associated with food insecurity risk, accounting for important child and family characteristics. The random components at the bottom of the table suggest that the odds of food insecurity do indeed vary across neighborhoods (random intercept) and, perhaps more importantly, the effect of family socioeconomic adversity varies across neighborhoods (random slope). Thus, we turn to asking whether neighborhood disadvantage moderates the impact of family socioeconomic adversity on food insecurity. Table 3 provides results for cross-level interactions that allow us to evaluate whether children in the most disadvantaged families face special heightened risks of food insecurity when they live in the most disadvantaged neighborhoods (Hypothesis 4) or whether a disadvantage paradox exists and these children actually have a lower risk of food insecurity (Hypothesis 5). Specifically, we provide results for an interaction between family socioeconomic adversity and neighborhood concentrated disadvantage. We also tested interactions between family socioeconomic adversity and the neighborhood density of large supermarkets, convenience stores, and social assistance organizations, but they were not significant so we do not present them here (full results available upon request). Table 3. Multilevel Logistic Regression Models for Household Food Insecurity: Interaction Models (Odds Ratios) Model 1 OR Model 2 OR Child and Family Characteristics Child is male 1.04 1.04 Child's age in months 1.01 1.01 Foreign-born parent 1.27** 1.28** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.82+ 0.82+  Hispanic 1.02 1.02  Other 0.88 0.89 Mother's age in years 1.01** 1.01** Family socioeconomic adversity (SEA) 2.35*** 2.35*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.15+  Other family type 0.79* 0.79* Number of siblings in the household 1.18*** 1.18*** Years lived in home 0.96* 0.96* Family receives food stamps 1.63*** 1.63*** Child in fair/poor health 1.59** 1.59** Primary caregiver likely depressed 2.90*** 2.90*** Primary caregiver in fair/poor health 1.56*** 1.56*** Primary caregiver has work limitations 1.52*** 1.52*** Neighborhood Characteristics Total Population (ln) 0.87* 0.87* Urban area 1.29** 1.33* Concentrated disadvantage 1.22*** 1.22*** Concentrated disadvantage*Family SEA 0.76*** 0.76*** Neighborhood resources (density per square mile) Social assistance organizations 0.99 Constant 0.06*** 0.06*** sd(random intercept) 0.30*** 0.30*** sd(random slope - Family SEA) 0.11* 0.11* Model 1 OR Model 2 OR Child and Family Characteristics Child is male 1.04 1.04 Child's age in months 1.01 1.01 Foreign-born parent 1.27** 1.28** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.82+ 0.82+  Hispanic 1.02 1.02  Other 0.88 0.89 Mother's age in years 1.01** 1.01** Family socioeconomic adversity (SEA) 2.35*** 2.35*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.15+  Other family type 0.79* 0.79* Number of siblings in the household 1.18*** 1.18*** Years lived in home 0.96* 0.96* Family receives food stamps 1.63*** 1.63*** Child in fair/poor health 1.59** 1.59** Primary caregiver likely depressed 2.90*** 2.90*** Primary caregiver in fair/poor health 1.56*** 1.56*** Primary caregiver has work limitations 1.52*** 1.52*** Neighborhood Characteristics Total Population (ln) 0.87* 0.87* Urban area 1.29** 1.33* Concentrated disadvantage 1.22*** 1.22*** Concentrated disadvantage*Family SEA 0.76*** 0.76*** Neighborhood resources (density per square mile) Social assistance organizations 0.99 Constant 0.06*** 0.06*** sd(random intercept) 0.30*** 0.30*** sd(random slope - Family SEA) 0.11* 0.11* *** < 0.001 ** < 0.01 * < 0.05 + < 0.10 Table 3. Multilevel Logistic Regression Models for Household Food Insecurity: Interaction Models (Odds Ratios) Model 1 OR Model 2 OR Child and Family Characteristics Child is male 1.04 1.04 Child's age in months 1.01 1.01 Foreign-born parent 1.27** 1.28** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.82+ 0.82+  Hispanic 1.02 1.02  Other 0.88 0.89 Mother's age in years 1.01** 1.01** Family socioeconomic adversity (SEA) 2.35*** 2.35*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.15+  Other family type 0.79* 0.79* Number of siblings in the household 1.18*** 1.18*** Years lived in home 0.96* 0.96* Family receives food stamps 1.63*** 1.63*** Child in fair/poor health 1.59** 1.59** Primary caregiver likely depressed 2.90*** 2.90*** Primary caregiver in fair/poor health 1.56*** 1.56*** Primary caregiver has work limitations 1.52*** 1.52*** Neighborhood Characteristics Total Population (ln) 0.87* 0.87* Urban area 1.29** 1.33* Concentrated disadvantage 1.22*** 1.22*** Concentrated disadvantage*Family SEA 0.76*** 0.76*** Neighborhood resources (density per square mile) Social assistance organizations 0.99 Constant 0.06*** 0.06*** sd(random intercept) 0.30*** 0.30*** sd(random slope - Family SEA) 0.11* 0.11* Model 1 OR Model 2 OR Child and Family Characteristics Child is male 1.04 1.04 Child's age in months 1.01 1.01 Foreign-born parent 1.27** 1.28** Child race/ethnicity (non-Hispanic white, ref)  Non-Hispanic black 0.82+ 0.82+  Hispanic 1.02 1.02  Other 0.88 0.89 Mother's age in years 1.01** 1.01** Family socioeconomic adversity (SEA) 2.35*** 2.35*** Family type (two biological parents, married or cohab., ref)  Single mother family 1.15+ 1.15+  Other family type 0.79* 0.79* Number of siblings in the household 1.18*** 1.18*** Years lived in home 0.96* 0.96* Family receives food stamps 1.63*** 1.63*** Child in fair/poor health 1.59** 1.59** Primary caregiver likely depressed 2.90*** 2.90*** Primary caregiver in fair/poor health 1.56*** 1.56*** Primary caregiver has work limitations 1.52*** 1.52*** Neighborhood Characteristics Total Population (ln) 0.87* 0.87* Urban area 1.29** 1.33* Concentrated disadvantage 1.22*** 1.22*** Concentrated disadvantage*Family SEA 0.76*** 0.76*** Neighborhood resources (density per square mile) Social assistance organizations 0.99 Constant 0.06*** 0.06*** sd(random intercept) 0.30*** 0.30*** sd(random slope - Family SEA) 0.11* 0.11* *** < 0.001 ** < 0.01 * < 0.05 + < 0.10 Turning to Model 1, we see that neighborhood concentrated disadvantage interacts with family socioeconomic adversity such that neighborhood disadvantage is associated with 22% higher odds of food insecurity for families at the mean level of socioeconomic adversity, but that this association diminishes as family socioeconomic adversity increases. In other words, families with more socioeconomic adversity have lower odds of food insecurity as neighborhood disadvantage increases. To illustrate, we generated predicted probabilities for food insecurity for three categories of family socioeconomic adversity: Mean Family socioeconomic adversity, Higher Family socioeconomic adversity (families at two standard deviations above the mean), and Lower Family socioeconomic adversity (families at two standard deviations below the mean). In Figure 1, we plot probabilities of food insecurity for the three groups by standardized neighborhood concentrated disadvantage. So for example, when neighborhood concentrated disadvantage equals zero (or is at the mean), families with low socioeconomic adversity have a probability of food insecurity of .03; families at the mean level of socioeconomic adversity have a probability of .10; and those with high socioeconomic adversity have a probability of .36. In general, the figure shows that the probability of food insecurity increases with neighborhood disadvantage for children in families at the mean level of socioeconomic adversity or lower.2 But for children in families at high levels of socioeconomic adversity, the probability of food insecurity dramatically decreases as neighborhood disadvantage increases, providing support for a disadvantage paradox hypothesis. Figure 1. View largeDownload slide Predicted Probability of Household Food Insecurity by Family Socioeconomic Adversity and Neighborhood Concentrated Disadvantage Figure 1. View largeDownload slide Predicted Probability of Household Food Insecurity by Family Socioeconomic Adversity and Neighborhood Concentrated Disadvantage Finally, in Model 2 of Table 3 we examined whether the cross-level interaction between neighborhood concentrated disadvantage and family socioeconomic adversity is robust after including the density of social assistance organizations. In other words, is lower risk for food insecurity for low SES families in higher disadvantaged neighborhoods due to greater access to social services in those neighborhoods? Accounting for social service organizations in Model 2 does not change the relationship between family socioeconomic adversity and neighborhood concentrated disadvantage. Put differently, this paradoxical relationship is not explained by accounting for a higher density of social assistance organizations in the neighborhood. DISCUSSION Scholars interested in poverty, collective action, and how communities impact individuals should be concerned with household food insecurity. With over 20% of households with children in the U.S. struggling to acquire enough food for a healthy and active lifestyle and signs that this problem is on the rise rather than receding, researchers, child advocates, policy makers, and the nation as a whole have a responsibility to better understand and provide solutions to curb and ultimately eliminate food insecurity. Our results demonstrate that food insecurity is not a simple or predictable measure of household disadvantage. Rather, food insecurity may be unevenly distributed even among the disadvantaged. Using nationally representative data, we clarify if and to what extent neighborhood characteristics may be influencing food insecurity in households that include young children. First, we investigated whether neighborhood characteristics predicted household food insecurity over and above individual and/or family characteristics. Similar to other recent studies (Kirkpatrick and Tarasuk 2010; Sadler, Gilliland and Arku 2013), we find that family, not neighborhood, characteristics are most directly linked with food insecurity. Indeed, families with young children in the U.S. facing social and economic disadvantages are struggling to provide enough food for their households. Alleviating poverty, ensuring equal access to educational and employment opportunities, and similar broad sweeping policy goals, can provide far reaching benefits for families with young children. But not all households in poverty are food insecure, and some non-poor households struggle with food security (Gundersen, Kreider and Pepper 2011). Thus, we posited that community characteristics might impact well-being differently for families at different levels of disadvantage. Indeed, existing evidence suggests that the impact of neighborhood disadvantage (Wodtke, Harding and Elwert 2016) and the ways in which communities might respond to adversity (Crowe and Smith 2012) are not monolithic. We presented and tested hypotheses on competing perspectives aimed at understanding how neighborhood and family characteristics impacted children in the most disadvantaged families – would family and neighborhood disadvantage accumulate to increase the risk of food insecurity, or would they interact in unexpected ways? Our results shed new light on the apparent null relationship between neighborhood disadvantage and household food insecurity. Increasing neighborhood disadvantage does associate with increasing risk of food insecurity for more well-off families. But for the most disadvantaged families, we find support for a disadvantage paradox – children in the most disadvantaged families and living in the most disadvantaged neighborhoods have lower probabilities of food insecurity than similar children in less disadvantaged neighborhoods. We investigated the possibility that this was a result of access to greater social resources, such as the placement of food pantries and other social organizations in the most disadvantaged neighborhoods, and find that these resources do little to explain the paradox. Rather, although we are unable to test this mechanism with our data, this finding may be a result of vulnerable residents taking on more informal strategies to deal with their own disadvantaged circumstances (Desmond 2012; Klinenberg 2002; Stack 1974). In all, these results suggest a contextual incongruence between family and neighborhood characteristics impacting the risk of food insecurity. In other words, when the level of family disadvantage is substantively different from the level of neighborhood disadvantage, resources that may be leveraged to mitigate food insecurity may be more difficult to access. Further, the neighborhood results suggest that social processes, and not necessarily physical access to food, are at the core of understanding how environment influences risk of food insecurity. Our null results regarding access to grocery stores are consistent with a growing body of work that fails to find robust associations between physical access to quality foods and well-being (Cummins, Flint and Matthews 2014; Sadler, Gilliland and Arku 2013). In addition, we find little evidence that increased access to formal social services influences risk, suggesting that families in need may engage in more informal ways to share information and obtain sufficient food supplies. Poor families in the most destitute communities may be employing effective strategies to share information and resources to counter food insecurity, particularly when it impacts children (Morton et al. 2005; Piff et al. 2010; Stack 1974). For example, dense, reciprocal ties among single mothers may be an important survival strategy in low-income communities (Dominguez and Watkins 2003), and participation in, not just the presence of, social organizations may be another survival strategy (Rankin and Quane 2000). This suggests that there may be close links among the poor which resemble those in mid-century American suburbs (Newman 1992), and that the meaning and use of social capital differs in high versus low SES neighborhoods (Altschuler, Somkin and Adler 2004; Sherman 2006). For example, residents of high SES neighborhoods may utilize social capital more to gain access to resources like job leads, but residents of low SES neighborhoods may use those ties more for daily survival (Desmond 2012). These social connections may be strongest among the most disadvantaged families in the most disadvantaged neighborhoods, possibly a product of moral capital and a closely shared set of strategies on how to cope with extreme adversity (Sherman 2006). Though not as often a research focus, scholars have documented coordinated strategies among residents in disadvantaged communities to overcome seemingly overwhelming hardship (Seccombe 2002). Indeed, prior evidence and the findings presented here support the idea that unique social processes are at work in disadvantaged communities. But more work is needed. Specifically, how might resource-sharing in poor communities work to reduce food insecurity and what sorts of policies can support those efforts? As a start, we might turn to existing policies, such as free and reduced breakfast and lunch programs in schools, and expand similar programs through community centers, especially when school is out of session. Such community food centers could purchase food in bulk, reducing cost, and provide education and training programs targeted to specific nutritional goals matched to the kinds of foods available. In addition, after-school programs which feed children (and sometimes their families) dinner could be expanded at schools with a higher proportion of students who receive free and reduced lunch. Such programs harness and nurture social capital and can address food insecurity at a community level. That children in less disadvantaged families in more disadvantaged neighborhoods do not receive the same reduction to food insecurity risk as children in the most disadvantaged families is interesting. In lower-income communities, social resources are mobilized to address chronic problems, and to gain access to resources residents cannot through financial means (Altschuler, Somkin and Adler 2004). These reciprocal exchanges of goods and services are strongest among residents of similar means (Dominguez and Watkins 2003; Menjivar 2000). Given these findings in conjunction with our own, we believe this is compelling evidence that less disadvantaged families may be more isolated in more disadvantaged communities, and less able (or less willing) to access the social networks of their more disadvantaged neighbors or social organizations which are designed to provide assistance. It is important to note that the ECLS-K is weighted toward children living in urban areas, and so some of these processes may differ for households with children in rural areas. Gaining a better understanding of the mechanisms underlying this “disadvantage paradox” will not only illuminate how community resources can be leveraged to alleviate food insecurity, but also how social ties and community resources are utilized by the poor to survive. The first and second authors acknowledge support from the University of Kentucky Center for Poverty Research (UKCPR) through funding from the Food and Nutrition Service of the Department of Agriculture (Contract No. AG-3198-S-12-0044) and the first author acknowledges support from the Young Scholars Program of the Foundation for Child Development (FCD) (Grant No. YSP Rice 10-2014). The opinions and conclusions expressed herein are solely those of the authors and should not be construed as representing the opinions or policies of the UKCPR, the FCD, or any agency of the Federal Government. The authors thank Laura Freeman for research assistance and the Kinder Institute Urban Health Program at Rice University for administrative support. Footnotes 1 Zip codes were converted to census tracts via crosswalk files generated from the Missouri Census Data Center’s MABLE/Geocorr2K application (http://mcdc2.missouri.edu/websas/geocorr2k.html). 2 Because the number of poor households in advantaged neighborhoods and the number of wealthy households in disadvantaged neighborhoods is relatively small, we re-estimated our models after eliminating outliers and found much the same results (available upon request). For presentation, the figure represents condensed results after eliminating outliers. REFERENCES Center for Law and Public Policy, (CLASP) . 2013 . Child Poverty in the U.S.: What New Census Data Tell Us About Our Youngest Children. Washington DC . Ahluwalia Indu B. , Dodds Janice M. , Baligh Magda . 1998 . “Social support and coping behaviors of low-income families experiencing food insufficiency in North Carolina.” Health Education & Behavior 25 5 : 599 - 612 . Alaimo K. , Briefel R. R. , Frongillo E. A. , Olson C. M. . 1998 . “Food insufficiency exists in the United States: Results from the third National Health and Nutrition Examination Survey (NHANES III).” American Journal of Public Health 88 3 : 419 - 26 . Alaimo K. , Olson C. M. , Frongillo E. A. . 2002 . “Family food insufficiency, but not low family income, is positively associated with dysthymia and suicide symptoms in adolescents.” J Nutr 132 4 : 719 - 25 . Alaimo K. , Olson C. M. , Frongillo E. A. Jr. 2001 . “Food insufficiency and American school-aged children's cognitive, academic, and psychosocial development.” Pediatrics 108 1 : 44 - 53 . Alaimo K. , Olson C. M. , Frongillo E. A. Jr. , Briefel R. R. . 2001 . “Food insufficiency, family income, and health in US preschool and school-aged children.” Am J Public Health 91 5 : 781 - 6 . Allison P.D. 2002 . Missing Data : London : Sage University . Altschuler Andrea , Somkin Carol P. , Adler Nancy E. . 2004 . “Local services and amenities, neighborhood social capital, and health.” Social Science & Medicine 59 6 : 1219 - 29 . Anderson Elijah. 1990 . Streetwise: Race, Class, and Change in an Urban Community . Chicago, IL : University of Chicago Press . Anderson Elijah. 1999 . Code of the Street: Decency, Violence, and the Moral Life of the Inner City . New York : W.W. Norton & Company . Bell Bethany A. , Ferron John M. , Kromrey Jeffrey D. . 2008 . “Cluster Size in Multilevel Models: The Impact of Sparse Data Structures on Point and Interval Estimates in Two-Level Models.” in Proceedings of the American Statistical Association, Section on Survey Research Methods . Alexandria, VA : American Statistical Association . Bickel Gary , Nord Mark , Price Cristofer , Hamilton William , Cook John . 2000 . “Guide to Measuring Household food security, revised 2000.” Alexandria, VA : United States Department of Agriculture, Food and Nutrition Service . Borman Geoffrey D. , Dowling Maritza . 2010 . “Schools and Inequality: A Multilevel Analysis of Coleman's Equality of Educational Opportunity Data ” Teachers College Record 112 5 : 1201 - 46 . Carpiano R.M. , Lloyd J.E.V. , Hertzman C. . 2009 . “Concentrated affluence, concentrated disadvantage, and children's readiness for school: A population-based, multi-level investigation.” Social Science & Medicine 69 3 : 420 - 32 . Carpiano Richard M. 2008 . “Actual or potential neighborhood resources and access to them: Testing hypotheses of social capital for the health of female caregivers.” Social Science & Medicine 67 4 : 568 - 82 . Casey P. H. , Szeto K. , Lensing S. , Bogle M. , Weber J. . 2001 . “Children in food-insufficient, low-income families: prevalence, health, and nutrition status.” Arch Pediatr Adolesc Med 155 4 : 508 - 14 . Coleman-Jensen Alisha , Gregory Christian , Singh Anita . 2014 . “Household Food Security in the United States in 2013: Statistical Supplement.” Economic Research Service U.S. Department of Agriculture . Cook J. T. , Frank D. A. , Berkowitz C. , Black M. M. , Casey P. H. , Cutts D. B. , Meyers A. F. , Zaldivar N. , Skalicky A. , Levenson S. , Heeren T. , Nord M. . 2004 . “Food insecurity is associated with adverse health outcomes among human infants and toddlers.” J Nutr 134 6 : 1432 - 8 . Crowe Jessica , Smith Justin . 2012 . “The influence of community capital toward a community's capacity to respond to food insecurity.” Community Development 43 2 : 169 - 86 . Cummins Steven , Flint Ellen , Matthews Stephen A. . 2014 . “New Neighborhood Grocery Store Increased Awareness Of Food Access But Did Not Alter Dietary Habits Or Obesity.” Health Affairs 33 2 : 283 - 91 . De Marco M. , Thorburn S. . 2009 . “The relationship between income and food insecurity among Oregon residents: does social support matter?” Public Health Nutr 12 11 : 2104 - 12 . Desmond Matthew. 2012 . “Disposable Ties and the Urban Poor.” American Journal of Sociology 117 5 : 1295 - 335 . Dominguez Silvia , Watkins Celeste . 2003 . “Creating networks for survival and mobility: Social capital among African-American and Latin-American low-income mothers.” Social Problems 50 1 : 111 - 35 . Fram Maryah Stella , Frongillo Edward A. , Draper Carrie L. , Fishbein Eliza M. . 2013 . “Development and Validation of a Child Report Assessment of Child Food Insecurity and Comparison to Parent Report Assessment.” Journal of Hunger & Environmental Nutrition 8 2 : 128 - 45 . Fram Maryah Stella , Frongillo Edward A. , Jones Sonya J. , Williams Roger C. , Burke Michael P. , DeLoach Kendra P. , Blake Christine E. . 2011 . “Children are aware of food insecurity and take responsibility for managing food resources.” The Journal of Nutrition 141 6 : 1114 - 19 . Gundersen C. , Kreider B. , Pepper J. . 2011 . “The Economics of Food Insecurity in the United States.” Applied Economic Perspectives and Policy 33 3 : 281 - 303 . Guo Guang , Zhao Hongxin . 2000 . “Multilevel Modeling for Binary Data.” Annual Review of Sociology 26 : 441 - 62 . Halpern-Meekin Sarah , Edin Kathryn , Tach Laura , Sykes Jennifer . 2015 . It's Not Like I'm Poor: How Working Families Make Ends Meet in a Post-welfare World . Oakland, CA : Univ of California Press . Hung Chanjin , Myers Samuel L. . 1999 . “Do the poor pay more for food? An analysis of grocery store availability and food price disparities.” The Journal of Consumer Affairs 33 2 : 276 - 96 . Jencks Christopher , Mayer Susan E. . 1990 . “The Social Consequences of Growing Up in a Poor Neighborhood.” Pp. 111 - 85 in Inner-City Poverty in the United States , edited by Lynn Laurence E. , McGeary Michael G. H. . Washington, D.C .: National Academy Press . Jetter Karen M. , Cassady Diana L. . 2006 . “The availability and cost of healthier food alternatives.” American Journal of Preventive Medicine 30 1 : 38 - 44 . Jones S. J. , Jahns L. , Laraia B. A. , Haughton B. . 2003 . “Lower risk of overweight in school-aged food insecure girls who participate in food assistance: results from the panel study of income dynamics child development supplement.” Arch Pediatr Adolesc Med 157 8 : 780 - 4 . Jyoti D. F. , Frongillo E. A. , Jones S. J. . 2005 . “Food insecurity affects school children's academic performance, weight gain, and social skills.” Journal of Nutrition 135 12 : 2831 - 39 . Kalil A. , Chen J. H. . 2008 . “Mothers' citizenship status and household food insecurity among low-income children of immigrants.” New Dir Child Adolesc Dev 2008 121 : 43 - 62 . Kawachi Ichiro. 2010 . “Social Capital and Health.” Pp. 18 - 32 in Handbook of Medical Sociology , edited by Bird C.E. , Conrad P. , Fremont A.M. , Timmermans Stefan . Nashville, TN : Vanderbilt University Press . Kawachi Ichiro , Berkman Lisa F. (Eds.). 2003 . Neighborhoods and Health . New York : Oxford University Press . Kimbro Rachel Tolbert , Denney Justin T . 2015 . “Transitions Into Food Insecurity Associated With Behavioral Problems And Worse Overall Health Among Children.” Health Affairs 34 11 : 1949 - 55 . Kimbro Rachel Tolbert , Denney Justin T. , Panchang Sarita . 2012 . “Individual, Family, and Neighborhood Characteristics and Children’s Food Insecurity.” Journal of Applied Research on Children 3 1 :Article 8. Kirkpatrick S.I. , McIntyre L. , Potestio M.L. . 2010 . “Child hunger and long-term adverse consequences for health.” Archives of Pediatrics and Adolescent Medicine 164 8 : 754 . Kirkpatrick Sharon I. , Tarasuk Valerie . 2010 . “Assessing the Relevance of Neighbourhood Characteristics to the Household Food Security of Low-Income Toronto Families.” Public Health Nutrition 13 7 : 1139 - 48 . Kirkpatrick Sharon I. , Tarasuk Valerie 2011 . “Housing Circumstances are Associated with Household Food Access Among Low-Income Urban Families.” Journal of Urban Health 88 2 : 284 - 96 . Kleinman R. E. , Murphy J. M. , Little M. , Pagano M. , Wehler C. A. , Regal K. , Jellinek M. S. . 1998 . “Hunger in children in the United States: potential behavioral and emotional correlates.” Pediatrics 101 1 : E3 . Klinenberg Eric. 2002 . Heat Wave: A Social Autopsy of Disaster in Chicago . Chicago, IL : The University of Chicago Press . Leventhal T. , Brooks-Gunn J. . 2000 . “The neighborhoods they live in: The effects of neighborhood residence on child and adolescent outcomes.” Psychological Bulletin 126 2 : 309 - 37 . Lichter Daniel T , Brown David L . 2011 . “Rural America in an urban society: Changing spatial and social boundaries.” Annual Review of Sociology 37 : 565 - 92 . Martin Katie S. , Rogers Beatrice L. , Cook John T. , Joseph Hugh M. . 2004 . “Social capital is associated with decreased risk of hunger.” Social Science & Medicine 58 12 : 2645 - 54 . McIntyre Lynn , Williams Jeanne V. A. , Lavorato Dina H. , Patten Scott . 2013 . “Depression and suicide ideation in late adolescence and early adulthood are an outcome of child hunger.” Journal of Affective Disorders 150 1 : 123 - 29 . Menjivar Cecilia. 2000 . Fragmented ties: Salvadoran immigrant networks in America . Berkeley, CA : University of California Press . Morenoff J. D. 2003 . “Neighborhood mechanisms and the spatial dynamics of birth weight.” American Journal of Sociology 108 5 : 976 - 1017 . Morland Kimberly , Diez Roux Ana V. , Wing Steve . 2006 . “Supermarkets, other food stores, and obesity: the atherosclerosis risk in communities study.” American Journal of Preventive Medicine 30 4 : 333 - 39 . Morton Lois Wright , Bitto Ella Annette , Oakland Mary Jane , Sand Mary . 2005 . “Solving the Problems of Iowa Food Deserts: Food Insecurity and Civic Structure.” Rural Sociology 70 1 : 94 - 112 . Morton Lois Wright , Bitto Ella Annette , Oakland Mary Jane , Sand Mary 2008 . “Accessing food resources: Rural and urban patterns of giving and getting food.” Agriculture and Human Values 25 1 : 107 - 19 . Murphy J. M. , Wehler C. A. , Pagano M. E. , Little M. , Kleinman R. E. , Jellinek M. S. . 1998 . “Relationship between hunger and psychosocial functioning in low-income American children.” J Am Acad Child Adolesc Psychiatry 37 2 : 163 - 70 . Newman Katherine S. 1992 . “Culture and structure in the truly disadvantaged.” City & Society 6 1 : 3 - 25 . Nord Mark , Coleman-Jensen Alisha , Andrews Margaret , Carlson Steven . 2010 . “Household food security in the United States, 2009.” Economic Research Report . Washington D.C .: US Department of Agriculture . Nord Mark , Hanson Karla . 2012 . “Adult caregiver reports of adolescents' food security do not agree well with adolescents' own reports.” Journal of Hunger & Environmental Nutrition 7 4 : 363 - 80 . Piff P.K. , Kraus M.W. , Côté S. , Cheng B.H. , Keltner D. . 2010 . “Having less, giving more: The influence of social class on prosocial behavior.” Journal of Personality and Social Psychology 99 5 : 771 . Rabe-Hesketh Sophia, , Skrondal Anders . 2008 . Multilevel and Longitudinal Modeling Using Stata . College Station, TX : StataCorp . Radloff Lenore Sawyer. 1977 . “The CES-D Scale: A Self-Report Depression Scale for Research in the General Population.” Applied Psychological Measurement 1 : 385 - 401 . Rankin Bruce H. , Quane James M. . 2000 . “Neighborhood poverty and the social isolation of inner-city African American families.” Social Forces 79 1 : 139 - 64 . Raudenbush Stephen W. , Bryk Anthony S. . 2002 . Hierarchical Linear Models . Thousand Oaks, CA : Sage Publications . Rose D. , Richards R. . 2004 . “Food store access and household fruit and vegetable use among participants in the US Food Stamp Program.” Public Health Nutrition 7 8 : 1081 - 88 . Royston Patrick. 2005 . “Multiple imputation of missing values: update.” The Stata Journal 5 2 : 188 - 201 . Ryu Jeong-Hee , Bartfeld Judith S. . 2012 . “Household food insecurity during childhood and subsequent health status: the Early Childhood Longitudinal Study--Kindergarten Cohort.” American Journal of Public Health 102 11 : e50 - e55 . Sadler Richard C. , Gilliland Jason A. , Arku Godwin . 2013 . “A Food Retail-Based Intervention on Food Security and Consumption.” International journal of environmental research and public health 10 8 : 3325 - 46 . Sampson Robert J. 2012 . Great American city: Chicago and the enduring neighborhood effect . Chicago, IL : University of Chicago Press . Sampson Robert J. , Raudenbush Stephen W. , Earls Felton . 1997 . “Neighborhoods and Violent Crime: A Multilevel Study of Collective Efficacy.” Science 277 5328 : 918 - 24 . Seccombe Karen. 2002 . ““Beating the odds” versus “changing the odds”: Poverty, resilience, and family policy.” Journal of Marriage and Family 64 2 : 384 - 94 . Sherman Jennifer. 2006 . “Coping with rural poverty: Economic survival and moral capital in rural America.” Social Forces 85 2 : 891 - 913 . Small Mario Luis. 2004 . Villa Victoria: The transformation of social capital in a Boston barrio : University of Chicago Press . Small Mario Luis. 2009 . Unanticipated gains: Origins of network inequality in everyday life . New York, NY : Oxford University Press . Small Mario Luis , Harding David J. , Lamont Michele . 2010 . “Reconsidering culture and poverty.” ANNALS, AAPSS 629 : 6 - 26 . Smith Sandra Susan. 2007 . Lone pursuit: Distrust and defensive individualism among the black poor . New York : Russell Sage Foundation . Stack C.B. 1974 . All our kin: Strategies for survival in a black community . New York : Basic Books . StataCorp . 2012 . “Stata Statistical Software: Release 13.” College Station, TX : Stata Corporation . Swaroop Sapna , Morenoff Jeffrey D. . 2006 . “Building community: The neighborhood context of social organization.” Social Forces 84 3 : 1665 - 95 . Van Hook J. , Balistreri K.S. . 2006 . “Ineligible parents, eligible children: Food Stamps receipt, allotments, and food insecurity among children of immigrants.” Social Science Research 35 1 : 228 - 51 . Vozoris Nicholas T. , Tarasuk Valerie S. . 2003 . “Household food insufficiency is associated with poorer health.” The Journal of Nutrition 133 1 : 120 -2 6 . Whitaker R. C. , Phillips S. M. , Orzol S. M. . 2006 . “Food insecurity and the risks of depression and anxiety in mothers and behavior problems in their preschool-aged children.” Pediatrics 118 3 : e859 - 68 . Wight Vanessa R. , Thampi Kalyani , Briggs Jodie . 2010 . “Who Are America's Poor Children? Examining Food Insecurity Among Children in the United States.” Columbia University, Mailman School of Public Health : National Center for Children in Poverty . Wilson William Julius. 1987 . The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy . Chicago, IL : University of Chicago Press . Wilson William Julius. 1996 . When Work Disappears: The World of the Urban Poor . New York : Alfred Knopf . Wodtke Geoffrey T. , Harding David J. , Elwert Relix . 2016 . “Neighborhood Effect Heterogeneity by Family Income and Developmental Period.” American Journal of Sociology 121 4 : 1168 - 222 . © The Author 2017. Published by Oxford University Press on behalf of the Society for the Study of Social Problems. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Social ProblemsOxford University Press

Published: Aug 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off