Deconstructing neighborhood effects across aggravated, domestic, and simple assault

ABSTRACT

Although prior research has long examined neighborhood effects on violent crime generally and aggravated assault specifically, less attention has been paid to how structure affects less serious offenses. However, neighborhood factors such as single-parent households or poverty are likely important predictors for the commission of less serious forms of violence, though their relative effects may vary depending on the level of severity of violence employed. The current study disaggregates assaults and examines whether neighborhood factors have differential effects on aggravated, non-aggravated (simple), and domestic assaults. Using data from Cleveland, OH we conduct a series of negative binomial regression models and use seemingly unrelated postestimation commands to identify differences across assault type. We identify important differences in the neighborhood drivers of assault, especially single-parent households, disadvantage, and residential stability, and between domestic assault from simple or aggravated. Our findings underscore the importance of disaggregating assaults to unmask distinctions in how community context influences serious and less serious forms of violence. In doing so, we identify important elements that crime reduction strategies should consider – particularly for less serious but more common forms of violence – in order to implemented more effectively.

KEYWORDS:

• Aggravated assault
• non-aggravated assault
• domestic assault
• neighborhood effects
• disadvantage
• residential stability

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. We interpret the terms ‘simple’ and ‘non-aggravated’ assault to mean the same thing and use them interchangeably throughout the manuscript.

2. Although some research has investigated the role of a subculture of violence associated with other racial/ethnic groups (e.g., Cao, Adams, and Jensen 1997; St. Vil et al. 2018) and in other contexts (e.g., Hackney 1969), the subculture of violence is largely assumed to be an urban Black phenomenon (Anderson 1999).

3. NEOCANDO aggregated individual incident reports to the census tract. Approximately 90% to 95% of all incidents were successfully coded to the tract level where the crime occurred. Incidents where a report was not filed or that were determined ‘unfounded’ are excluded. More information on the NEOCANDO crime data can be found here: https://neocando.case.edu/resources/neocando/new%20docs/08%20-%20Cleveland%20Police%20Data%20Indicators.pdf

4. The City of Cleveland lost population between 2000 and 2010 which meant that the Census consolidated some tracts together in 2010 that were separate in 2000. Therefore, to maintain the larger sample size, we use 2000 boundaries instead of harmonizing the crime data to 2010 (N = 174 tracts).

5. Note, in supplemental analyses we combined non-aggravated domestic and non-domestic non-aggravated assaults into one outcome to align with the aggravated assault count. The pattern of results was similar and the same variables were identified as significantly different between aggravated and non-aggravated simple assaults. However, because we also uncover meaningful differences between simple and domestic assault, we made the decision to keep the non-aggravated assaults disaggregated.

6. In ancillary models we included separate measures of the different components of the disadvantage index. We found a similar pattern of results, though the individual variables were not consistently significant across models. We decided to use the disadvantage index because it largely tells the same story and avoids unnecessary collinearity.

7. We also tested alternative models with the disaggregated components of the stability index. The different components of the index were consistently significant and negatively related to each assault type. We decided to include stability as an index to minimize multicollinearity and because there is no empirical or theoretical reason to believe that these factors would affect assaults differentially.

8. In alternative models we included the percent of single parent households in the disadvantage index. Doing so did not substantively alter the pattern or magnitude of the results and masked the independent effects of single parents on domestic assault specifically. For this reason and theoretical expectations, we kept the single parents variable separate from the disadvantage index.

9. We do not graph the predicted values for domestic assaults because the interaction is not a significant predictor of domestic assaults nor is it significantly different from one of the other assault types.

10. Though notably, the effect was not meaningfully different from the non-significant effect of disadvantage on aggravated or simple assaults shown in Table 2.

11. In alternative models we included an interaction term to determine whether the effect of stability on domestic assaults was contingent on the percent of single parent households, but we found no evidence of this to be the case.

Additional information

Notes on contributors

Lyndsay N. Boggess

Lyndsay N. Boggess is an Associate Professor in the Department of Criminology at the University of South Florida. Her research focuses on communities and crime, primarily how crime affects and is affected by neighborhood change, racial/ethnic composition, and gentrification and economic investment. Her work has appeared in Criminology, the Journal of Quantitative Criminology, and Crime & Delinquency.

Alyssa W. Chamberlain

Alyssa W. Chamberlain is an Associate Professor in the School of Criminology and Criminal Justice at Arizona State University. Her research focuses on neighborhood dynamics related to changes in social and demographic structure, housing, and inequality and the spatial distribution of crime. She also examines issues related to prisoner reentry and recidivism, and how incarceration and reentry both affect and are affected by community structure.

Lexi Gill

Lexi Gill is a doctoral student in the Department of Criminology at the University of South Florida. She earned her Master of Science in Criminal Justice and Criminology from the University of Missouri - Kansas City. She is a graduate research assistant on a Smart Policing Initiative grant. Her research interests include communities and crime, policing, and social network analysis.