Substantial heritability of neighborhood disadvantage: Individuals themselves might potentially contribute to a self-selection process that explains which neighborhoods they occupy as adults

Understanding neighborhood disadvantage: A behavior genetic analysis. Albert J. Ksinan, Alexander T.Vazsonyi. Journal of Criminal Justice, Volume 73, March–April 2021, 101782. https://doi.org/10.1016/j.jcrimjus.2021.101782

Abstract

Purpose Studies have shown that disadvantaged neighborhoods are associated with higher levels of crime and delinquent behaviors. Existing explanations do not adequately address how individuals select neighborhood. Thus, the current study employed a genetically-informed design to test whether living in a disadvantaged neighborhood might be partly explained by individual characteristics, including self-control and cognitive ability.

Method A sibling subsample of N = 1573 Add Health siblings living away from their parents at Wave 4 was used in twin analyses to assess genetic and environmental effects on neighborhood disadvantage. To evaluate which individual-level variables might longitudinally predict neighborhood disadvantage, a sample of N = 12,405 individuals was used.

Results Findings provided evidence of significant heritability (32%) of neighborhood disadvantage. In addition, a significant negative effect by adolescent cognitive ability on neighborhood disadvantage 14 years later was observed (β = −0.04, p = .002). Follow-up analyses showed a genetic effect on the association between cognitive ability and neighborhood disadvantage.

Conclusions Study findings indicate substantial heritability of neighborhood disadvantage, showing that individuals themselves might potentially contribute to a self-selection process that explains which neighborhoods they occupy as adults.

Introduction

Criminologists have extensively focused on the impact of neighborhood social disorganization on crime and deviance since the first half of the 20th century (Shaw & McKay, 1942). Research has provided evidence that neighborhoods with disorganized structural characteristics, including high levels of mobility, high rates of poverty, or high numbers of single-parent families, were associated with higher levels of criminal behavior (Bursik & Grasmick, 1999; Morenoff, Sampson, & Raudenbush, 2001; Sampson, 1985; Sampson, Raudenbush, & Earls, 1997; Wilson, 1987).

These hypothesized neighborhood effects have generally been considered to flow in one direction, namely from neighborhoods to individuals. However, a small number of studies have hypothesized and tested the opposite, namely that individuals select into their neighborhoods. Given that neighborhood variables reflect the aggregation of the qualities and characteristics of individual members, it seems likely that certain individual traits might predict neighborhood characteristics (Hedman & van Ham, 2012). If individual traits do in fact predict neighborhood characteristics and all psychological traits are to a certain extent heritable (Turkheimer, 2000), then it stands to reason that neighborhood characteristics will show some heritable effect as well. The current study used a genetically-informed design to test for both genetic and environmental effects on selecting into certain neighborhoods and to test whether individual characteristics (self-control and cognitive ability) have developmental effects on this selection process.

A neighborhood is defined as a geographically unique subsection or area, part of a larger community. Typically, neighborhoods are operationalized using geographic boundaries defined by an administrative agency (such as the Census Bureau), which partitions neighborhoods into tracts or blocks (Sampson, Morenoff, & Gannon-Rowley, 2002).

The traditional framework for studying neighborhood effects is rooted in social disorganization theory. According to this theory, every individual is prone to engage in some deviant or criminal behaviors. Bonds to society make these behaviors too costly and thus effectively prevent crime from happening. The neighborhood process through which it controls the behaviors of its members is termed collective efficacy (Morenoff et al., 2001) or the ability of individuals sharing a neighborhood to work together and to solve issues related to their neighborhood. In this way, individuals engage in effective indirect social control in order to prevent neighborhoods from deteriorating. A typical example of such indirect social control is when adults monitor youth loitering in the neighborhood and are willing to confront them when they disturb or disrupt a public space (Sampson et al., 1997). A well-functioning neighborhood is a complex and cohesive system of social networks, rooted in both the family as well as the community (Sampson, Morenoff, & Earls, 1999).

Neighborhood structural factors such as high poverty, single-parent families, residential instability, high unemployment, or a high number of minority inhabitants, are associated with lower levels of neighborhood organization or an inability of the community to maintain effective social control, according to social disorganization theory (Sampson, 1997; Sampson & Groves, 1989). The impact of these structural factors might lead to alienation of neighborhood members and low levels of investment in the community, which in turn leads to greater social disorder and thus higher proneness to disorder and crime (Leventhal & Brooks-Gunn, 2000; Leventhal, Dupéré, & Brooks-Gunn, 2009; Molnar, Miller, Azrael, & Buka, 2004; Sampson & Groves, 1989).

Empirical support for social disorganization theory and the concept of collective efficacy in predicting crime and delinquency has been provided by a number of studies that have used hierarchical or multi-level modeling. For example, Sampson et al. (1997) found that concentrated disadvantage, immigration concentration, and residential (in)stability significantly predicted collective efficacy, which in turn mediated the effects of disadvantage and residential (in)stability on several measures of violence. Similarly, Sampson and Raudenbush (1999) found that collective efficacy of a neighborhood predicted lower levels of disorder and crime (see also Molnar et al., 2004; Sampson, 1997; Valasik & Barton, 2017).

In contrast, a more recent approach to studying neighborhood effects has focused on neighborhood characteristics, including individual-level variables (as opposed to predicting rates in neighborhood). Based on Leventhal and Brooks-Gunn's review (2000), neighborhoods affect a plethora of individual adjustment measures. Among them, neighborhood SES was found to positively predict educational attainment, mental health, as well as negatively predict individual delinquency and criminal behavior (Leventhal et al., 2009).

Individuals do not randomly allocate into neighborhoods, but rather, they actively seek out and select their neighborhoods. If neighborhoods consist of individual members, it stands to reason that the likelihood of living in a particular place is, to some certain extent, affected by individual characteristics, and thus, that neighborhood characteristics are also affected by individual differences. This is referred to as ‘self-selection’. In the current definition, self-selection refers to a broader concept than simply ‘individuals making deliberate choices when deciding where to live.’ Such a view would be imprecise and potentially harmful, as it might put too much emphasis on personal responsibility for potentially detrimental living conditions. Rather, self-selection refers to a more impersonal process where individuals with different life histories occupy different life trajectories that lead them to different places of residence, and, in many cases, living in a particular neighborhood is not so much a volitional process or act, but rather a situation that cannot be easily changed.

The idea that a self-selection process might be taking place related to an association between an individual (or a family) and neighborhood characteristics is certainly not new. In fact, the issue with non-independence of neighborhood sorting and individual characteristics has been mentioned by several authors (Sampson & Sharkey, 2008). However, individual characteristics that were identified to influence self-sorting into particular neighborhoods were of a social nature, such as being a renter versus a homeowner, being single, or being an immigrant, just to name a few (Hedman & van Ham, 2012). At present, however, there does not appear to be a clear understanding about the potential effect of self-selection on neighborhood effects. Some research did not find support for neighborhood effect once self-selection was accounted for (Oreopoulos, 2003), while other studies found that neighborhood effects remained significant after accounting for self-selection (Aaronson, 1998; Dawkins, Shen, & Sanchez, 2005; Galster, Marcotte, Mandell, Wolman, & Augustine, 2007). Thus, the evidence is quite mixed.

Behavior genetic studies partition phenotypic variance into three sources: heritability, shared environment, and nonshared environment. Over the past three decades, studies have consistently shown both environmental and genetic influences on the vast majority of individual traits (Plomin, DeFries, Knopik, & Neiderheiser, 2013; Polderman et al., 2015). However, genetic effects are not limited to individual characteristics. In fact, some presumably environmental effects have also been found to be correlated with genetic predispositions. There are three types of gene-environment correlations: passive rGE, evocative rGE, and active rGE (Plomin, DeFries, & Loehlin, 1977). Particularly relevant to the concept of neighborhood self-selection is active rGE, which refers to individuals actively selecting environments based on their inherent preferences (Moffitt, 2005).

Because individuals are not randomly selected for certain environments as much as they are active agents in selecting, modifying, and adapting to the environments, this process is affected by their individual characteristics, which themselves are substantially affected by heritable materials. A review of 55 studies by Kendler and Baker (2007) showed that there are substantial genetic effects (average h2 = 0.27) on measures of the environment, including parenting behaviors, stressful life events, social support, or peer interactions. Nevertheless, there has not been a study that has directly tested the heritability of neighborhood characteristics. Most genetically-informed studies on more distal environmental effects (such as schools or neighborhoods) focused on their moderating effects only (Cleveland, 2003; Rowe, Almeida, & Jacobson, 1999). For example, a study by Connolly (2014) found that neighborhood disadvantage moderated the genetic effect on adolescent delinquency between the ages of 6 and 13 years, and between 14 and 17 years, where greater heritable effects were observed at higher levels of neighborhood disadvantage.

How might individual characteristics be genetically related to the neighborhoods that individuals live in? The key to understanding potential genetic effects on neighborhoods lies in the process of active rGE, according to which individuals actively ‘select’ their environments. In the case of neighborhoods, the selection process is both the selection of a particular neighborhood to live in as well as the variety of neighborhoods that are available, also determined to a certain extent by individual traits.

Neighborhood socioeconomic status is defined as socioeconomic status of individual houses or their inhabitants, and, in the context of the United States, socioeconomic status is strongly affected by the level of education, which in turn has been found to be positively associated with cognitive ability or intelligence (L. Gottfredson, 1997a; Neisser et al., 1996; Strenze, 2007). Differences in intelligence have a large heritable component which has been found to increase with age (Bouchard, Lykken, McGue, Segal, & Tellegen, 1990; Devlin & Daniels, 1997; Haworth et al., 2010). Moreover, a more direct link between cognitive ability or intelligence and career success, as well as intelligence and more positive developmental adjustment outcomes in general, was also established by numerous studies (Caspi, Wright, Moffitt, & Silva, 1998; L. Gottfredson, 2004; Judge, Higgins, Thoresen, & Barrick, 1999; Schmidt & Hunter, 2004). Thus, it stands to reason that neighborhood socioeconomic status should have a heritable or genetic component, and individual cognitive ability might partially explain this variance.

Another candidate personality trait, which might play a significant role in affecting neighborhood characteristics, is self-control or the ability to exercise restraint in delaying immediate gratification and subduing our impulses. Perhaps the most prominent theory emphasizing the role of self-control is self-control theory by Gottfredson and Hirschi (1990). According to Gottfredson and Hirschi, all deviant and criminal behaviors are to some extent related to a lack of self-control. A great number of studies have provided consistent empirical support that (low) self-control is perhaps the single best predictor of deviant and criminal behaviors (Hay, 2001; Vazsonyi, Mikuška, & Kelley, 2017; Wright, Caspi, Moffitt, & Silva, 1999), as well as better health, better career prospects, or less substance use (Casey et al., 2011; Mischel et al., 2011; Moffitt et al., 2011). In this view, the association between neighborhood disorganization and low self-control would consider low self-control as the cause rather than the outcome, as individuals with low self-control would be more likely to self-select into neighborhoods with higher levels of social disorganization (Caspi, Taylor, Moffitt, & Plomin, 2000; Evans, Cullen, Burton Jr., & Dunaway, 1997). Both cognitive ability and (low) self-control have in fact been tested in a longitudinal study by Savolainen, Mason, Lyyra, Pulkkinen, and Kokko (2017); findings showed that childhood differences in cognitive skills as well as childhood antisocial propensity (both measured at age 8) were traits that significantly foretold the developmental cascade which led to higher socioeconomic exclusion in midlife.