Stuck or Rooted? Perspectives on the Residential Immobility of Children in the U.S. from Poor Neighborhoods and Implications for Policy

Abstract

Families relocating from concentrated poverty neighborhoods is somewhat rare, either due to structural constraints that limit mobility or the disincentive to leave dense social networks built over time. Researchers previously juxtaposed these two experiences as either “stuck” or “rooted”. We advance a critical take on both perspectives by demonstrating the heterogeneity of life in disadvantaged neighborhoods for Black urban youth. We utilize data from the Panel Study of Income Dynamics and the 1997 Child Development Supplement to investigate contextual immobility, barriers to moving, and self-reported levels of neighborhood social ties to critique prior research and emergent policy that categorizes disadvantaged populations as “stuck” or “rooted”. Our findings demonstrate that immobility is most strongly associated with the household head lacking a high school education and with knowing more children’s names in the neighborhood. Thus, immobility is associated with structural barriers to moving and social rootedness. We discuss how current policy strategies do not effectively address this duality. We conclude that policy strategies should facilitate intragenerational mobility through housing choice, including the choice to remain in the neighborhood.

1. Introduction

Chronic exposure to neighborhood poverty can be problematic for children’s mental and physical health, education, and future economic mobility (Burdick-Will et al. 2011; Sharkey 2016; Xue et al. 2005). Moreover, families living in poor neighborhoods generally tend to live in such places for long periods. Although they may move around within poor neighborhoods, mobility out of poor neighborhoods is rare and becomes more unlikely the longer a family lives in poverty (Chetty et al. 2020; Wilson 1987). Because of the negative impacts and persistence of poverty over time, it is common for researchers and policymakers to portray children and families chronically living in poor neighborhoods as “stuck” (Sharkey 2013).

However, the experience of growing up in a poor neighborhood is not uniform. Children and families living in poor neighborhoods experience a range of social environments and often possess many social assets (Dahl et al. 2010). Their social networks are often dense and close-knit (Stack 1974), and they are more likely than households in more affluent neighborhoods to have family members living nearby (Ackert et al. 2019). Residents of poor neighborhoods also frequently give and receive support and share resources with nearby friends and family (Jarrett et al. 2010). These social dynamics suggest that many children and families persistently living in poor neighborhoods could also be described as “rooted”.

In this study, we review the prior literature on residential immobility from poor neighborhoods through Hunter and Robinson’s (2016) deficit and asset framework. In doing so, we identify “stuck” and “rooted” as two prevalent and competing discourses surrounding children and families living in concentrated poverty. We then conduct an analysis focusing on Black children living in urban areas, describing how our study population can be characterized as both “stuck” and “rooted”. Our focus on children stems from prior research on the intergenerational transmission of poverty (Chetty and Hendren 2018; Sharkey and Elwert 2011), which highlights how exposure to impoverished neighborhoods in childhood produces lasting impacts over the life course (Alvarado 2018; Crowder and South 2011; Wodtke et al. 2011). Our focus on Black children is motivated by our desire to build on prior research, including the influential works of Wilson (1987), Massey and Denton (1993), Sharkey (2013), Pattillo (1999, 2007), and others that have emphasized the dynamics of Black urban neighborhoods.

Our goal is to add to these discussions by demonstrating the heterogeneity of life in poor neighborhoods and the duality of poor populations being both “stuck” and “rooted”. We utilize longitudinal data from the Panel Study of Income Dynamics to track residential (im)mobility, barriers to moving, and self-reported levels of neighborhood social ties. We use these descriptive findings as a jumping-off point to critique prior research and policy that portrays poor populations as “stuck” or “rooted” rather than both simultaneously. As we will discuss, not only does the characterization of low-income families as “stuck” versus “rooted” conjure differing images of these families; it also suggests vastly different pathways for addressing the problem of persistent poverty.

2. Background and Theory

2.1. Poverty among Black Urban Youth

Over the past seven decades, child poverty in the United States has fluctuated between 14 and 27% (U.S. Bureau of the Census 2021). Over this time, Black youth have consistently had roughly three to four times the poverty rates of white youth (U.S. Bureau of the Census 2021). Black youth are also more likely than white youth to be persistently poor. Nearly 40% of Black children spend more than half of their childhood in poverty, compared to 5% of white children (Ratcliffe and McKernan 2012). These racial disparities have remained steady over time (Ratcliffe and McKernan 2012).

Poverty rates differ substantially across races and ethnicities but also by neighborhood. Poverty rates rose in urban areas in the 1970s and 1980s and again in the 2000s, with poverty clustering along geographic lines (Kneebone and Nadeau 2015). Growth in the concentration of poverty has exacerbated economic and racial segregation and social isolation (Kneebone and Nadeau 2015). Analysis of recent trends indicates that growing income segregation and weak economic performance have increased the likelihood of poor people living in high-poverty neighborhoods, with poor Black people being at the greatest risk (Iceland and Hernandez 2017).

The implications of poverty early in life are vast. Childhood poverty has been linked to chronic stress, high school non-completion, and teenage childbearing (Evans and Schamberg 2009; Ratcliffe and McKernan 2012). Adults who grew up in poverty are less consistently employed than those who did not and have worse mental and physical health outcomes (Ratcliffe and McKernan 2012). Concentrated neighborhood poverty is associated with a similar set of adverse outcomes. Above and beyond a child’s personal poverty status, neighborhood poverty is linked to high school non-completion, adult joblessness, and reduced income in adulthood (Alvarado 2018; Crowder and South 2011; Wodtke et al. 2011). Given the well-established negative impacts of childhood poverty, a focus of U.S. poverty policy has been on residential mobility out of poor neighborhoods (Ludwig et al. 2013). However, little is known about children and families who remain in poor neighborhoods and what drives the outcome to stay, despite there being much research on concentrated poverty. The research, as we will show, tells conflicting stories about families remaining in poor neighborhoods depending on the frame through which it views residential immobility.

2.2. Theoretical Framework

This study examines residential immobility from poor neighborhoods through two broad frames established by Hunter and Robinson (2016): the deficit frame and the asset frame. Given the well-established negative impacts of childhood poverty, it is unsurprising that most poverty scholarship operates through a deficit frame. According to Hunter and Robinson (2016, p. 385), the deficit frame emphasizes the “structures that negatively affect Black urban life (e.g., disappearance of work, residential segregation, poor education, urban poverty) and the cultural “deficits” that either are adaptations to those structural realities or (as some deficit scholars argue) are the cause of urban Black hardships”. When applied to poverty scholarship, the deficit frame focuses on how concentrated poverty generally goes hand-in-hand with under-resourced schools, increased crime, and deteriorated built environments (Sampson 2009). In turn, according to the deficit perspective, these contextual disadvantages make it difficult for children to “escape” poor neighborhoods (Sharkey 2013).

Hunter and Robinson (2016) offer a corollary approach called the asset frame. The asset frame focuses on “the agency and cultural contributions of urban Black Americans” and “the actions and attitudes of urban Black Americans as consequential to the making and unmaking of places, organizations, and institutions and policies” (Hunter and Robinson 2016, pp. 385, 387). When applied to childhood poverty, the asset frame emphasizes how children and families living in poor areas forge meaningful local ties and place attachments. These ties are often imperative to surviving day-to-day in poverty (Garrett-Peters and Burton 2016; Jarrett 1995; Stack 1974; Burton and Clark 2005). Residents’ relationships with their neighborhoods through emotional significance, community, and culture can take the form of metaphorical rootedness (Gustafson 2001; Malkki 1992) and influence decisions to move from the neighborhood.

Building on the deficit and asset frames, we outline two perspectives for how scholars approach childhood poverty research: the stuck and the rooted perspectives. Comparing and contrasting the stuck and rooted perspectives reveals they are two fundamentally different approaches to understanding persistent residence in poor neighborhoods, and each suggests a very different course of policy action.

2.3. The “Stuck” Perspective

By focusing extensively on the harmful outcomes of life in poor neighborhoods and the structural barriers to moving, the literature on concentrated poverty emphasizes a deficit frame. The Moynihan Report drew national attention to the plight of poor Black residents of urban neighborhoods and the negative effects of concentrated poverty (Moynihan 1965). The report focused on the “disorganization” of Black families and their communities as the cause of chronic poverty. Spurred by a conservative political agenda, the culture of poverty proponents latched on to the family disorganization narrative, failing to recognize the systemic conditions that created racialized neighborhood inequality (Cherry 1995; Roach and Gursslin 1967; Ryan 1971). Instead, the culture of poverty arguments blamed residents of poor inner-city Black neighborhoods for their circumstances, citing culture and behavior as reasons for concentrated poverty.

Wilson (1987), in response to the culture of poverty narratives, suggested that several structural factors drove concentrated racial poverty in urban environments. First, was the widespread joblessness of Black men caused by changes to U.S. trade policy and technological advances that eliminated many well-paying industrial jobs in urban cores. Consequently, according to Wilson (1987) and others (Kasarda 1989; Wacquant and Wilson 1989), joblessness led to higher rates of poverty and lower rates of marriage in Black urban neighborhoods. Second, as Black middle-class residents abandoned the inner city, suburbanization increased socioeconomic inequality and class segregation in the Black population (Wilson 1987; Cullen and Levitt 1999). In addition, inner-city institutions such as churches and schools suffered without the support of the Black middle class, and young people who were left behind lacked access to role models and social network opportunities, contributing to their social isolation (Katzman 1983; Wilson 1987, 1996).

Beyond a focus on the economic forces that created the conditions to trap poor Black people in impoverished urban neighborhoods, the prior literature also emphasizes the role of structural racism. Massey and Denton (1989) are credited with bringing a focus to racism’s role in the “hyper-segregation” of poor inner-city Black populations. They and others suggest that racial discrimination and intimidation have been effective tools to confine Black households to particular areas (Massey and Denton 1989; Bell 2008; Roscigno et al. 2009). Scholars argue that white people intentionally created the black ghetto to isolate and control the growing population of Black people living in the city (Massey and Denton 1989; Logan et al. 2015). This line of research also advanced the idea that Black inner-city residents developed cultural characteristics that increased their marginalization and decreased their chances of succeeding in mainstream society (Massey and Denton 1989; Wilson 1987). The concept of an “underclass” became a powerful image and narrative to describe the Black urban poor.

Other research emphasizing a deficit frame has focused on the intergenerational transmission of poverty. For example, Sharkey (2013, p. 133) argued that residents of poor urban neighborhoods experience a “legacy of disadvantages” across generations, ranging from health outcomes to school performance. Researchers have emphasized how intergenerational poverty limits the chances of residential and economic mobility, leading Sharkey (2013) and others (Corcoran 1995; Vartanian et al. 2007) to characterize poor families as “stuck in place”. The Moving to Opportunity (MTO) experiment was a large-scale test of whether housing vouchers could lead to residential mobility for poor households, with mixed results (Briggs et al. 2010). When this line of research acknowledges social connections that bind people to poor neighborhoods, it is often to point out how social ties to poor areas increase the exposure of Black people of all incomes to problems such as crime, substandard schools, and other structural inequities (Sharkey 2013; Pattillo 2007; Sampson and Morenoff 2004).

On the other hand, the stuck discourse illuminates systemic stratification by focusing on the structural explanations of concentrated poverty. This perspective also tends to focus on the deficits of poor Black urban neighborhoods. The stuck discourse implies one reason that people continue to live in poor Black neighborhoods is that it is overwhelmingly difficult to leave. Scholars such as Wilson (1987), Massey and Denton (1993), and Sharkey (2013) were instrumental in refuting the culture of poverty arguments that blame individuals for systemic social problems. However, by approaching poor Black urban neighborhoods from a deficit perspective, these and other scholars have largely ignored the agency and value of community that poor Black people living in urban environments may experience (Hunter and Robinson 2016). This could be in part due to these scholars’ desire to maintain a structuralist perspective on poverty, or it could reflect the data at hand. Neighborhood life is generally poorly captured in large-scale quantitative censuses and surveys such as those utilized by Wilson, Massey, Denton, and Sharkey, which may explain why the stuck narrative developed without much acknowledgment of the value and meaning some residents attach to their, albeit poor, neighborhoods.

2.4. The “Rooted” Perspective

Even as the stuck perspective was being developed and reinforced, a counternarrative emerged that described urban Black communities from an asset frame. Recognizing that individuals living in poverty may experience constrained networks due to concentrated poverty, restricted geographical mobility, and less social capital (Coleman 1988; Fischer 1982), this research demonstrates that social ties still benefit the urban Black poor, particularly the youth (Domínguez and Watkins 2003; Feagin 1970, p. 11970; Oliver 1988; Stack 1974). Much of this research argues that due to absent “weak ties” to societal institutions, individuals develop strong ties to family members and neighbors out of necessity (Granovetter 1973). Although restricted social morphology can further disadvantage individuals in chronic neighborhood poverty, their social connections are vital to maintaining community (Anguelovski 2013; Klinenberg 1999) and getting by in daily life (Stack 1974). Thus, within the urban Black community, there is support for what Wellman and Leighton (1979) described as the “community-saved” argument. The community-saved argument states that structural forces such as industrialization and urbanization necessitate communal ties that rely on mutual aid and reciprocity (Oliver 1988; Wellman and Leighton 1979). Moreover, we can tie the community-saved argument to Black people being rooted in neighborhoods experiencing chronic poverty.

In the literature, Black communities are viewed through an asset frame that reveals how social ties can root individuals to poor neighborhoods in various ways. Low-income individuals tend to live closer to family members than high-income individuals, suggesting that poor individuals may experience a particularly tight social geography (Ackert et al. 2019). Social ties are known to discourage individuals from moving and play an important part in selecting a destination neighborhood among movers (Spring et al. 2017). Social connections can also help explain why some participants who received a housing voucher through the MTO experiment chose to move back to poor neighborhoods after living briefly in higher-opportunity areas. The return movers often had strong social ties to their former neighborhoods (Briggs et al. 2010; Quillian 2012; Schoenbaum 2017). Thus, the nature and geography of social ties might play an important role in binding residents to poor neighborhoods.

The social norm of mutual aid also serves to root individuals in disadvantaged neighborhoods. The idea of collaboration and the reciprocity of goods and services among both fictive and blood kin recur as themes in the literature (Aschenbrenner 1975; Bentelspacher et al. 2006; Johnson 1999; Shimkin and Shimkin 1974; Stack 1974). Stack (1974) described an intricate exchange system within Black social networks based on favors that provided the necessary support for daily life. Mutual aid was such a strong norm that it was frowned upon to turn down requests. Once individuals had received a favor, they were obligated to carry out one in return, either right away or at some point in the future (Stack 1974). An obligation to this mutual aid structure can be one more force rooting people to their neighborhoods.

In addition, the feelings of knowing the neighborhood and of being known can root people to their neighborhood. The comfort and familiarity that people experience in their neighborhoods create feelings of place attachment and deter residential mobility (Shelby 2017). Place attachment can also encompass historical connections such as churches, community groups, or family history (Lewicka 2010; Low and Altman 1992). Due to this attachment, long-term neighborhood residents often have a different outlook on chronic neighborhood poverty (Dahl et al. 2010). Residents often view outsiders as having skewed perceptions of their neighborhoods since outsiders are not cognizant of how these communities function and view chronically poor neighborhoods as socially disorganized (Shelby 2017). Residents are committed to combating a dysfunctional view through cooperation that can take multiple forms, such as activism, civic engagement, and other types of collective action (Anguelovski 2013; Saegert 2006; Shelby 2017; Weffer 2017).

To some, there is a blanket assumption that individuals who are experiencing or affected by poverty are all poor, which is not always true. This belief has highlighted a diversity of experiences for individuals staying or coming back to neighborhoods in chronic poverty. For example, Pattillo (2007) found that the Black middle class often takes on a “protector” role to advocate for their lower-income neighbors. Although the Black middle class is evidence of upwards social mobility, they are still Black and are affected by the same racial inequalities that Black lower-class individuals face (Pattillo 2007). Despite their class difference, the Black middle class may feel like they belong and have a collective responsibility to better their neighborhoods because of their class status.

An asset frame also brings to light the agency families use to manage their children’s lives in poor neighborhoods (Jarrett 1995). Some seek out after-school programs and other opportunities that place children in safe settings and provide opportunities for mobility (Furstenberg 1993). Others closely monitor their child’s activities in the neighborhood and plan their family’s activities within known safe areas or safe times of the day (Jarrett 1997). Families also use collective sharing of resources and intensive and supportive ties to local kin to help buffer their children against the negative aspects of their environment (Jarrett et al. 2010). These studies demonstrate the power of rootedness as a productive strategy to manage the hardships of concentrated poverty. They also highlight the heterogeneity of life in poor neighborhoods. Indeed, reflecting on his 20-year longitudinal study of teenage mothers and their offspring, Furstenberg (1993, p. 231) noted that “differences…may be as conspicuous and consequential as any commonalities” within this group.

2.5. Linking the Stuck and Rooted Perspectives

Applying the asset and deficit framework to residential immobility reveals how urban Black families living in chronically poor neighborhoods have been understood as stuck and rooted to varying degrees. Combining the discourses may be productive for understanding and ameliorating the problem of persistent childhood poverty. Structural systems of inequality drastically impact what neighborhoods are available, accessible, and affordable to Black families. Yet, within (or due to) these constraints, Black families can develop meaningful and productive community ties that in turn make moving away less likely. Recent scholarship has begun to note these endogenous dynamics. Krysan and Crowder’s (2017) social structural sorting perspective describes racial segregation as a self-fulling prophecy, whereby racial segregation leads to segregated social networks, producing racially-disparate knowledge about and connections to specific neighborhoods. In work on Black placemaking, Hunter and colleagues describe how Black people have exerted spatial agency to push back against structural racism (Hunter et al. 2016; Hunter and Robinson 2016). According to Ewing (2018, p. 471), Hunter and colleagues’ research serves as an “intentional counterpoint to the presumption that black people can only passively receive the space allotted to them through a mixture of white supremacy and happenstance; in this framing, black people are agents who shape and define the space around them”. Each of these perspectives demonstrates the interlocking nature of structure and agency.

We apply similar logic to our conceptualization of stuck and rooted dynamics. We ask, to what extent is residential immobility in poor neighborhoods impacted by stuckness and rootedness simultaneously? Utilizing moves from a high-poverty to a low-poverty neighborhood as our dependent variable (i.e., “contextual mobility”), we investigate associations between moving versus staying with stuck indicators, such as high school non-completion, unemployment, and teenage childbearing, and rooted indicators, such as knowing the names of people in the neighborhood and having friends and family in the neighborhood. By exploring relationships between contextual (im)mobility, barriers to moving, and local social connections, our research seeks to disentangle and complicate the discourse surrounding children and families who remain in poor neighborhoods. Complicating the discourse, we believe, is key to developing more effective anti-poverty urban policy.

3. Materials and Methods

3.1. Data

Our study focuses on a nationally-representative sample drawn from the U.S. and focusing on Black youth living in U.S. metropolitan areas. Drawing on data from the Panel Study of Income Dynamics (PSID) and decennial census tract data, we explore intragenerational mobility by investigating children, their caregivers, and their contextual environments over time. The PSID is a longitudinal survey of U.S. residents and their families that began in 1968. For our analysis, we selected children who participated in the 1997 Child Development Supplement (CDS). The CDS is a follow-up survey that supplemented the 1997 PSID by gathering more detailed data about the socio-demographic, psychological, and economic aspects of children from the main sample (Hofferth et al. 1997). The CDS could randomly select up to two children who were 12 and under within a household. All PSID families were eligible, and the response rate was 88% (Hofferth et al. 1997). Although the CDS was drawn from the main PSID sample, which originally oversampled low-income families, it is still representative of our study population.

The primary caregiver, defined as an adult who lives with the child and provides the majority of their care, answered the survey questions utilized in our analysis. The primary caregiver was the mother 94% of the time, but it could also be the father or another legal guardian (PSID 2006). The 1997 CDS sample consisted of 3563 children ranging from 0 to 12 years old (Hofferth et al. 1997). We excluded 1330 children that did not have a CDS household questionnaire completed by the primary caregiver. Of the remaining children, we focused on those who lived in a metropolitan area in 1997 and whose household head was Black (n = 791). We used the supplemental Geospatial Match Files to link the addresses of sample children to corresponding codes for their census tracts in 1997 and two years later, in 1999. We then attached tract poverty data from the 1990 and 2000 decennial censuses and estimated 1997 and 1999 tract poverty rates with linear interpolation. We selected only children living in high-poverty tracts in 1997, defined as a tract poverty rate at or above 20%. The final analytic sample was 460 children.

The data time period was selected for several reasons. First, the 1997 CDS survey collected several indicators of social rootedness in neighborhoods that have not been collected by the PSID since. The 1997 CDS data is unique since very few nationally representative surveys include information on neighborhood social ties. Second, the 1997 CDS aligns with the timing of seminal research on contextual mobility from the stuck perspective, and in particular, the MTO experiment, which began in 1994. Our study uses data from the same era but incorporates novel measures of social rootedness.

3.2. Key Measures

3.2.1. Contextual (Im)mobility

Our outcome of interest is intragenerational contextual mobility. Sharkey (2013, p. 17) defined contextual mobility as the change in residential location that encompasses what children “see and hear and breathe and experience around them when they engage in public life outside the home, along with the political, economic, and social forces that affect individuals on the basis of where they live”. Our study sample consisted of children living in high-poverty neighborhoods in 1997. To operationalize contextual mobility, we identified the poverty rate of the children’s neighborhoods at the subsequent PSID interview, two years later. We then created a binary indicator where a value of 1 indicated that the child moved to a low-poverty neighborhood, and 0 indicated that the child remained in a high-poverty neighborhood. Following prior research in this area (Galster et al. 2003; Ren and Morrow-Jones 2014), we used a poverty rate of 20% as the cut-off between low-poverty and high-poverty neighborhoods. Some children who remained in high-poverty neighborhoods may still have moved, but they moved between high-poverty neighborhoods. Thus, contextual immobility is not the same as never moving because some children move around frequently without ever exiting neighborhood poverty.

3.2.2. Family Barriers to Moving

We operationalized the stuck perspective with family-level characteristics that likely serve as barriers to exiting high-poverty neighborhoods. These indicators were drawn from prior research identifying hardships and outcomes associated with living in concentrated poverty (Alvarado 2018; Crowder and South 2011; Ratcliffe and McKernan 2012; Wodtke et al. 2011). They include whether, in 1997, the household head was unemployed, had not completed high school, and was not married. Whether the child’s mother was a teenager at birth and whether the family’s 1997 income was below the federal poverty threshold were also included.

3.2.3. Neighborhood Social Ties

We operationalized the rooted perspective with five questions from the 1997 CDS that captured neighborhood social ties. The child’s primary caregiver answered these questions, which included: How difficult is it for you to tell a stranger in your neighborhood from someone who is a resident? (0 = very difficult to 2 = not at all difficult). How many of the adults living in your neighborhood do you talk with regularly? How many children or teenagers living in your neighborhood do you know by name? Not counting family members who live with you, how many family members live in your neighborhood? How many good friends do you have that live in your neighborhood? Respondents answered with the actual number (except for the first question). Descriptive analysis revealed respondents’ answers to the family question were highly skewed. To reduce skewness, we top-coded responses to this question to 20 family members. We explored other strategies to deal with skewness, including top-coding at different thresholds, top-coding the other continuous measures, which were also skewed but not as extreme as family members, and logging the measures. We found no substantive differences in the results, so only family members were top-coded.

Correlations between our measures are a potential concern since we might expect respondents to report high levels of social ties across many or all of the neighborhood social ties variables. However, correlations between the measures were moderate, and follow-up analysis of variance inflation factors following regression models found no evidence of multicollinearity. Because respondents self-defined the neighborhood they were “rooted” in (rather than administrative or interviewer-provided definitions of neighborhoods), we included a control for the respondents’ definitions of their neighborhoods in our analysis.

3.2.4. Covariates

To account for the heterogeneity of experiences within poor neighborhoods, we also measured a host of family and household characteristics drawn from prior research (Clampet–Lundquist 2010; Guest et al. 2006; South et al. 2005). Household size, composition, housing tenure, and housing type are shown to influence whether families move and how people interact with and relate to their neighbors. A child’s age and sex may also be influential on their primary caregivers’ perspective on the neighborhood. We therefore measured several covariates, including the number of children under 18 in the household, the age of the household head, the age of the child, the child’s sex, whether the family owned their home, whether the family lived in public housing, and their length of residence in the same home in years. We also included a binary indicator of living in the South since most of our sample (61%) were Southern residents. We also adjusted for the primary caregiver’s self-reported definition of their neighborhood, coded as 0 = the block or street you live on, 1 = the block or street and several streets, 2 = an area within a 15-min walk, and 3 = an area larger than a 15-min walk. All covariates were measured in 1997, prior to a potential move.

3.3. Analytical Strategy

Our analysis aims to illuminate the contribution of stuck and rooted indicators to contextual (im)mobility. We begin by exploring descriptive statistics, first for the full sample, and then separately for those who were contextually mobile versus those who were contextually immobile. We then explore associations between stuck and rooted indicators and contextual mobility in a multivariate framework. Next, we estimate logistic regressions predicting contextual mobility, first with stuck indicators, then with rooted indicators, and then with both simultaneously plus the individual and household covariates. A small number of variables had missing values for some respondents. We use multiple imputations with chained equations to fill in missing data. We use a total of 10 imputed datasets, with imputation models predicting missing values with all key measures and covariates from our analysis, following White et al. (2011). We also incorporate clustered standard errors, clustered by family ID, to account for the non-independence of observations for siblings in our sample. Our results establish an association but not a causal relationship between contextual mobility and stuck and rooted indicators. We discuss the implications of these associations for our understanding of intragenerational contextual mobility.

4. Results

4.1. Sample Characteristics

Table 1. Sample Characteristics.

	Mean/%	SD	Median	Mode
Contextual mobility
Move out of a high-poverty (>20% poor) neighborhood by next interview (1 = yes)	18.48%
Indicators of the stuck perspective
HH unemployed	13.70%
HH did not complete high school	33.70%
HH not married	63.48%
Mother under age 20 at birth	21.30%
Family income below poverty threshold	51.96%
Indicators of the rooted perspective
Ability to identify strangers in neighborhood
Very difficult	17.39%
Somewhat difficult	29.13%
Not at all difficult	53.48%
Family members in neighborhood	8.98	8.42	6	0
Adults you talk with in neighborhood	5.06	10.28	3	0
Children you know by name in neighborhood	17.84	29.77	7	0
Good friends in neighborhood	3.36	5.78	2	0
Individual and family characteristics
Number of children < 18 in household	2.62	1.39	2	2
HH age	36.49	10.51	35	35
Child’s age	6.56	3.62	6	1
Child’s sex (1 = female)	45.65%
Family owns home	32.39%
Family lives in public housing	23.91%
Length of residence in years	6.33	5.86	4	14
Region of residence (1 = South)	61.30%
Definition of neighborhood
Block or street you live on	36.30%
Block or streets and several streets	29.57%
Area within 15 min walk	17.39%
Area larger than 15 min walk	16.74%
N	460

HH = Household head; First of 10 imputation data sets.

presents descriptive statistics for key measures within our study population of children with a Black head of household living in a metropolitan area. Statistics for contextual mobility suggest a high level of continued exposure to poor neighborhoods, with approximately 18% of children moving out of a high poverty neighborhood by the subsequent interview. Irrespective of contextual mobility, indicators of the stuck perspective show that 13.7% of our sample had an unemployed household head, and around one-third of household heads did not complete high school. Most of our sample, 63%, lived in households where the head was not married. Approximately 21% of sample children were born to mothers under the age of 20. Roughly half of the families within our sample were below the poverty threshold, underscoring that not all families living in high-poverty neighborhoods are themselves poor.

Indicators of the rooted perspective show fairly high levels of social ties to the neighborhood. Approximately 53% of primary caregivers found it not at all difficult to identify strangers in their neighborhoods. About one-third found it somewhat difficult, whereas 17% found it very difficult. Based on the medians, most primary caregivers regularly talked to three adults, knew seven children by name, and had six family members and two good friends in their neighborhood. However, there was substantial variation around these numbers, with many primary caregivers reporting no friends or family in the neighborhood and knowing or talking with no one. Other primary caregivers reported very high numbers, which skewed the means upwards. Some of the variations could be due to respondents self-defining their neighborhoods. Approximately 36% defined their neighborhood as the block or street they lived on, 30% as the block or streets and several streets, 17% as an area within a 15-min walk, and 16% as an area larger than a 15-min walk. Values for social ties reported here are similar to those reported by Logan and Spitze (1994), who used data from the same era but on a different sample. In their study, people knew 12.6 neighbors well enough to call them by name and had 11 family members living in the same metropolitan area, on average. The authors also noted the distributions of their social ties variables were highly skewed (Logan and Spitze 1994).

Additional characteristics revealed that the average age of children in our sample was 6.5 years old, and 45% of children identified as female. The average age of the household head was approximately 36. On average, families had two children under 18 in the household and mainly lived in the South (61%). Roughly 32% of children lived in owner-occupied homes, and about 24% lived in public housing. Overall, our sample aligns with other descriptions of urban Black children in the U.S. (Drake and Rank 2009; Plybon and Kliewer 2001).

4.2. Sample Characteristics by Contextual (Im)mobility

In

Table 2. Sample Characteristics by Contextual Mobility.

	Whether Moved Out of a High-Poverty Neighborhood
	Movers		Stayers
	Mean/%	SD	Mean/%	SD
Indicators of the stuck perspective
HH unemployed	11.76%		14.13%
HH did not complete high school	20.00%		36.80%		**
HH not married	54.12%		65.60%		*
Mother under age 20 at birth	23.53%		20.53%
Family income below poverty threshold	42.35%		54.13%		*
Indicators of the rooted perspective
Ability to identify strangers in neighborhood
Very difficult	11.76%		18.67%
Somewhat difficult	28.24%		29.33%
Not at all difficult	60.00%		52.00%
Family members in neighborhood	9.59	8.70	8.93	8.39
Adults you talk with in neighborhood	5.97	15.44	4.85	8.72
Children you know by name in neighborhood	12.58	24.94	19.03	30.69	*
Good friends in neighborhood	3.60	8.41	3.30	5.01
Individual and family characteristics
Number of children < 18 in household	2.34	1.17	2.68	1.43	*
HH age	34.62	9.83	36.92	10.62	*
Child’s age	5.48	3.56	6.81	3.59	**
Child’s sex (1 = female)	45.88%		45.60%
Family owns home	31.76%		32.53%
Family lives in public housing	23.53%		24.00%
Length of residence in years	6.61	6.19	6.26	5.79
Region of residence (1 = South)	68.24%		59.73%
Definition of neighborhood
Block or street you live on	38.82%		35.73%
Block or streets and several streets	18.82%		32.00%
Area within 15 min walk	21.18%		16.53%
Area larger than 15 min walk	21.18%		15.73%
N	85			375

HH = Household head; * Significantly different from movers (p < 0.05); ** Significantly different from movers (p < 0.01); First of 10 imputation data sets.

, we divide the sample into children who moved out of a high-poverty neighborhood by the next interview (i.e., “movers”) and children who remained in a high-poverty neighborhood (i.e., “stayers”). Descriptive analysis revealed some significant differences between these groups. One of the largest differences between movers and stayers was the high school completion of the household head. More than 36% of stayers had a household head that did not complete high school, compared to 20% of movers. In addition, 65% of stayers lived in a household where the head was not married, compared to 54% of movers. Furthermore, 54% of stayers had a family income below the poverty level, compared to 42% of movers. Stayers’ primary caregivers knew significantly more children by name in the neighborhood than movers’ primary caregivers. Stayers knew a mean of 19 children, while movers knew a mean of 12.5. However, movers and stayers did not significantly differ on other measures of social ties to the neighborhood.

There were also small but significant differences in the number of children in the household, the household head’s age, and the child’s age for movers versus stayers. Stayers were slightly older (in terms of the child’s and the household head’s ages) and had slightly more children in the household. On many other characteristics, movers and stayers were similar. Stayers were no more likely to own their home or live in public housing than movers, and both had similar lengths of residence in the neighborhood. With our preceding analyses showing bivariate associations between contextual mobility and some key characteristics, we next looked at multivariate estimates of exiting high-poverty neighborhoods.

4.3. Multivariate Analysis

Table 3. Logistic Estimates of Exiting High-Poverty Neighborhoods for Black Youth Living in Metropolitan Areas.

	Model 1			Model 2			Model 3
	or		(se)	or		(se)	or		(se)
Indicators of the stuck perspective
HH unemployed	1.099		(0.512)				0.875		(0.422)
HH did not complete high school	0.460	*	(0.170)				0.412	*	(0.157)
HH not married	0.672		(0.238)				0.579		(0.219)
Mother under age 20 at birth	1.366		(0.465)				1.093		(0.411)
Family income below poverty threshold	0.784		(0.280)				0.678		(0.282)
Indicators of the rooted perspective
Ability to identify strangers in NH (ref = very difficult)
Somewhat difficult				1.632		(0.820)	1.508		(0.806)
Not at all difficult				2.028		(0.993)	1.661		(0.899)
Family members in NH				1.017		(0.020)	1.019		(0.023)
Adults you talk with in NH				1.022		(0.018)	1.024		(0.019)
Children you know by name in NH				0.976	*	(0.010)	0.978	*	(0.010)
Good friends in NH				1.030		(0.030)	1.035		(0.035)
Individual and family characteristics
Number of children < 18 in household							0.948		(0.122)
HH age							0.979		(0.020)
Child’s age							0.902	*	(0.041)
Child’s sex (1 = female)							1.088		(0.277)
Family owns home							0.657		(0.330)
Family lives in public housing							0.912		(0.410)
Length of residence in years							1.018		(0.032)
Region of residence (1 = South)							1.459		(0.577)
Definition of neighborhood (ref = block or street you live on)
Block or streets and several streets							0.613		(0.271)
Area within 15 min walk							1.241		(0.660)
Area larger than 15 min walk							1.394		(0.635)
Constant	0.373	**	(0.112)	0.128	***	(0.062)	1.086		(0.962)
N	460			460			460

* p < 0.05; ** p < 0.01; *** p < 0.001; Combined results of 10 imputation data sets. HH = household head. NH = neighborhood.

presents our multivariate analysis where we estimated associations between stuck and rooted indicators and contextual mobility. We have shown evidence of associations in our descriptive analysis, but here we examined the robustness of those associations in a logistic model accounting for additional factors. We estimated three separate models: one predicting contextual mobility with stuck indicators, one with rooted indicators, and one with both accounting for additional covariates.

Model 1 demonstrates stuck indicators’ associations with contextual mobility, accounting for no other factors. Although indicators in the model were moderately correlated, a follow-up analysis of variance inflation factors indicated minimal evidence of multicollinearity. The results indicate that high school non-completion by the household head was significantly related to contextual mobility. Children whose household heads did not complete high school had 54% lower odds of exiting a high-poverty neighborhood than children whose household heads completed high school. Other indicators of the stuck perspective were not significantly related to contextual mobility. These results suggest that high school non-completion of the household head is a particularly substantial barrier to contextual mobility for Black children living in high-poverty neighborhoods.

Model 2 investigates social ties to neighborhoods and their associations with contextual mobility. Again, evidence of multicollinearity in the follow-up analysis was minimal, allowing the various rooted indicators to be modeled together. The results revealed that adults knowing children in the neighborhood significantly impacted the likelihood of contextual mobility. Each additional child known in the neighborhood was associated with a 2.4% decrease in the odds of exiting a high-poverty neighborhood. Caregivers may be hesitant to move their child from a neighborhood where they already know many other children. These families may be enmeshed in the local schools or involved in watching each other’s children. They may also feel a sense of safety in knowing the names of their child’s playmates. Other types of social ties to neighborhoods were not significantly related to contextual mobility in this sample.

Model 3 demonstrates the independent associations of the stuck and rooted indicators, accounting for each other and for additional covariates. When accounting for other factors, the results for high school non-completion and knowing children in the neighborhood do not change. High school non-completion of the household head significantly reduced the odds of contextual mobility, irrespective of social ties to the neighborhood and other factors. In other words, the reduction in mobility associated with high school non-completion is not due to families with lower education levels being more socially rooted in their neighborhoods. Caregivers knowing children by name in the neighborhood also reduced the odds of contextual mobility despite any stuck barriers to moving and other factors. A child’s age was also associated with contextual mobility, so each additional increase of one year in age was associated with a 9.8% decrease in the odds of exiting a high-poverty neighborhood.

Consistent with our descriptive results, we found evidence that contextual mobility was associated with high school completion, knowing children by name, and the age of children in the household. The multivariate analysis demonstrates the robustness of those associations, which we interpret as evidence that children and families can be independently stuck and rooted at the same time. Below, we discuss the pitfalls of compartmentalizing contextual mobility into the stuck or rooted perspectives, as well as options and opportunities for individuals living in chronic poverty to benefit from established connections to their neighborhoods.

5. Discussion

This paper set out to demonstrate the interconnections between barriers to residential mobility and experiences of neighborhood social ties within poor neighborhoods, or what we juxtaposed as being “stuck” and being “rooted”. We grounded this analysis in the deficit and asset frameworks (Hunter and Robinson 2016) with the goal of establishing how these frameworks have structured scholars’ understandings of contextual immobility. The deficit perspective, in particular, often manifests in a tendency to see mobility as preferable and immobility as indicative of individual or household deficits that stem from structural inequalities. This mobility bias ignores factors rooting people in their neighborhoods. We aimed to clarify the relationship between stuck and rooted indicators and contextual mobility, given that they are likely also related to each other.

Our findings demonstrated that Black urban youth could be stuck and rooted simultaneously, complicating notions of residential immobility from poor neighborhoods. Our analyses suggested that high school non-completion is a substantial barrier to exiting high-poverty neighborhoods. At the same time, families knowing children in the neighborhood is an important driver of staying. Overall, our results demonstrated a paradox for housing policy, whereby children who are seemingly stuck in high-poverty neighborhoods can also be socially rooted.

Our findings lead us to a reassessment of the existing anti-poverty housing policy. The stuck discourse has informed existing policy focusing on residential mobility as the pathway out of poverty. The stuck discourse implies policy remedies that prioritize breaking up segregated communities, increasing access to lower-poverty areas, and facilitating moves out of high-poverty neighborhoods. The housing choice voucher system, commonly referred to as “Section 8”, has been the major policy approach to meeting these directives. Other policy interventions include the Hope VI program and its predecessor, the Choice Neighborhoods Initiative, which have focused on replacing low-quality public housing with new mixed-income communities.

These programs have produced mixed results. Vouchers have helped some families access safe and affordable housing, but vouchers have been ineffective for other families (Rosen 2020). In addition to being difficult to obtain, vouchers can be difficult to use in some areas if landlords are unwilling to take them (Rosen 2020). Moreover, families who successfully use their vouchers to leave a poor neighborhood often move back over time, citing a desire to be close to their existing social support networks and places to which they are familiar (De Souza Briggs et al. 2010; Schoenbaum 2017). Under the Hope VI program, some former public housing residents have reported high satisfaction with their new units, but others have reported feeling isolated from their former communities and unwelcome or stigmatized living next to market-rate residents (McCormick et al. 2012). Critics see these programs as a systematic continuation of the forced removal of Black communities (Goetz 2015). They argue that programs focused on relocating the Black urban poor result in psychological impacts similar to the “root shock” experienced by plants when they are hastily moved (Fullilove 2016). Overall, these policy failures reflect the shortcomings of the stuck perspective, which largely ignores the everyday relationships and attachments rooting many poor people to their neighborhoods.

The rooted perspective implies different policy strategies than the stuck perspective. It starts from the premise that there are social assets present in poor neighborhoods that are worth preserving and building upon. The urban community development movement reflects such ideas by seeing the solution to racially concentrated areas of poverty as hyper-targeted quality of life interventions (Bhattacharyya 2004; Frisch and Servon 2006). Policy solutions include the Community Development Block Grant program, the Community Reinvestment Act, the Promise Neighborhoods program, and affordable housing development programs led by community development corporations operating at the neighborhood level (Galster et al. 2003; Wachter and Ding 2020). Each of these programs seeks to improve the quality of life within high-poverty neighborhoods rather than moving people out.

Yet there are important caveats to rooted policy approaches. The social strategies poor Black families use to manage the hardships of concentrated poverty do not solve poverty itself (Corcoran 1995; Duncan and Zuberi 2006; Greenbaum et al. 2008; Harper et al. 2003; Rankin and Quane 2000; Wagmiller and Adelman 2009). Relying on rooted social networks to address poverty is not sustainable long-term, as Black social networks end up overburdened and tapped out (Dahl et al. 2010; Duck 2012; Garrett-Peters and Burton 2016; Miller-Cribbs and Farber 2008). Place-based interventions thus require an influx of resources, but new resources must be accompanied by robust plans to maintain affordable housing. Otherwise, improvements to neighborhood infrastructure can easily induce gentrification (Davidson 2008; Layser 2019; Mehdipanah et al. 2018; Richardson et al. 2020). In addition, existing place-based policies do little to break up areas of concentrated white affluence, which many scholars see as the real root of the problem (Goetz et al. 2019).

Thus, with limited federal dollars to support low-income families, policymakers have a dilemma: should the federal government provide resources and programs to move low-income families out of low-income neighborhoods or work to improve the place itself? Our findings suggest that a more effective policy solution would do both. Poverty intervention needs to occur at a structural or systemic level, but at the same time, it cannot neglect micro-level realities. The development of a more effective anti-poverty urban policy begins by rejecting the oversimplification of the lived experience of Black children and families. The distillation of their experience, especially into the “stuck” narrative, delegitimizes human agency and overemphasizes the inevitability of the market (Harvey 1978; Vale 2013). A new solution would prioritize tangible policies that acknowledge the rooted-stuck dialectic by fostering and preserving productive social ties, neighborhood attachments, and individual choice while simultaneously addressing structural barriers to mobility.

More research is necessary before new policy solutions can take shape. We see at least three important avenues for future research. The first is to gain a better understanding of the relationship between barriers to mobility and social ties to neighborhoods. Research should explore whether increased neighborhood ties are an independent outcome of poverty exposure (e.g., perhaps poverty causes people to rely on their neighbors more, fostering greater neighborhood ties) or are simply correlated with it through some other factor. The second is to examine feedback loops between residential immobility and neighborhood ties. The longer a family lives in the neighborhood, the more chance they have had to establish social ties, but social ties may in turn lengthen their stay in the neighborhood. This research should also consider involuntary moves caused by evictions and other displacements, which prior research shows disrupt local social ties (Manzo et al. 2008). The third is to investigate how subjective versus objective neighborhood conditions influence residential mobility. Research focusing on neighborhood perceptions shows that people experience the same neighborhood differently, and that perceptions influence residential decisions (Jones and Dantzler 2021). How individuals perceive their social ties to neighbors, whether as positive, negative, beneficial, obligatory, or otherwise, could help clarify how neighborhood ties impact different families. Perceptions might vary by individual and family characteristics such as poverty and employment status, and family structure, so interactions between perceptions of social ties and other characteristics should also be explored.

To accomplish this research, scholars will need to utilize qualitative and quantitative approaches since each on its own yields a limited view. As a whole, the censuses and large-scale surveys utilized by quantitative researchers have provided important insights into poverty rates and changes over time, yet they have systematically failed to capture the rooted aspects of people’s lives and frequently oversimplify the experience of Black families. Policymakers should also look to qualitative research, where insights into how neighborhoods and social networks operate are more prominent. Doing so can lead to a reconceptualization of stuckness and rootedness that sees them as intersecting factors rather than polar opposites. Understanding where individual families fall along stuck and rooted dimensions may help explain why some families are well served by existing housing policies, and others are not.

Research elucidating stuck and rooted dynamics would also be useful for evaluating emerging policy solutions. For example, scholars have suggested fair housing law revisions and the expansion of social and legal support for voucher holders (Rosen 2020; Galster 2019; DeLuca 2019). Others have suggested building institutional capacity to administer housing programs at the local level by building partnerships between the federal government and local organizations (Galster 2019; Howell et al. 2019). Others have called for resident-led development or neighborhood efficacy interventions, especially those that mandate and strictly enforce affordable housing commitments (Immergluck 2022). The framework of stuckness and rootedness could help assess the impacts of these and other emerging policy solutions on heterogenous populations and communities. How effectively do such interventions connect individual and community goals, especially for families with complex stuck/rooted dynamics? Do such interventions facilitate intragenerational mobility though housing choice, including the choice to remain in the neighborhood?

The United States is entrenched in stratification and inequality. Although Black children and families living in poverty experience meaningful social relationships and survive through reciprocal exchange within their social networks, social networks are not a means to end poverty. The structural and systemic inequality in the U.S. that perpetuates chronic neighborhood poverty will persist unless there is an attempt to ameliorate social problems through policy. However, as Small et al. (2010) point out, that policy must be culturally aware and culturally informed.