Violent Crime Arrest Gap

Violent Crime Arrest Gap

Go Back to methodology

About

The Violent Crime Arrest Gap measures the ratio of violent crime-related arrests among Black Americans to violent crime-related arrests among White Americans relative to their respective populations in a county. Individuals from both Hispanic and non-Hispanic ethnicities are included. According to the Federal Bureau of Investigation’s (FBI) Uniform Crime Reporting (UCR) program, violent crimes include four offenses: murder and nonnegligent manslaughter, rape, robbery, and aggravated assault 1. In the U.S., Black individuals are arrested for violent crimes at a higher rate than White individuals. For instance, one study found Black individuals to be arrested for violent crimes at a rate 2.7 times higher than White individuals 2. 

Why is the Violent Crime Arrest Gap important to the Structural Racism and Discrimination (SRD) Index?

The Violent Crime Arrests Gap between the dominant White race and marginalized races, such as the Black race, is a critical outcome of the persistent racist and discriminatory laws and policies in the criminal justice system of the U.S. Racial and ethnic minorities, particularly Black individuals, are disproportionately arrested for violent crimes compared to their White counterparts. This disparity is often attributed to structural disadvantages and systemic biases present in law enforcement practices 3. The plausible drivers of violent crime pertain to socioeconomic disadvantages, such as income inequality, lack of quality education, food insecurity, and strained mental health services, which are often prevalent in segregated neighborhoods with higher concentrations of Black populations. Segregated neighborhoods often experience increased police surveillance and enforcement, leading to higher arrest rates for violent crimes among communities of color 4. It is this confluence of compounding racial and socioeconomic inequalities that appears to cultivate the cycles of community violence and hostile police-citizen relationships, which continue to plague Black communities living in disenfranchised neighborhoods 5. 

What is the expected relation to Structural Racism and Discrimination?

A higher value of the Violent Crime Arrest Gap between Black and White populations contributes to a higher value or score of the SRD Index.

How is the Violent Crime Arrest Gap Gap calculated?

Data Source

We obtained the number of arrests for Violent Crime reported to the FBI’s Uniform Crime Reporting (UCR) Program each year by police agencies in the U.S., summarized yearly from the Inter-university Consortium for Political and Social Research (ICPSR) of the University of Michigan.  We obtained the population data disaggregated by race from the IPUMS National Historical Geographic Information System (NHGIS) 6. All data are publicly available. Below are the specifics of the data files used from ICPSR.

2020: Uniform Crime Reporting Program Data: Arrests by Age, Sex, and Race, Summarized Yearly, United States, 2020 (ICPSR 38788).

2010: 1) Uniform Crime Reporting Program Data: Arrests by Age, Sex, and Race, Summarized Yearly, United States, 2010 (ICPSR 33522); 2) Uniform Crime Reporting Program Data: County-Level Detailed Arrest and Offense Data, United States, 2010 (ICPSR 33523).

2000: 1) Uniform Crime Reporting Program Data [United States]: Arrests by Age, Sex, and Race, Summarized Yearly, 2000 (ICPSR 3997); 2) Uniform Crime Reporting Program Data [United States]: County-Level Detailed Arrest and Offense Data, 2000 (ICPSR 3451).

1990: 1) Uniform Crime Reporting Program Data [United States]: Arrests by Age, Sex, and Race, Summarized Yearly, 1990 (ICPSR 23341); 2) Uniform Crime Reporting Program Data [United States]: County-Level Detailed Arrest and Offense Data, 1990 (ICPSR 9785).

For all years: Jacob Kaplan’s Concatenated Files: Uniform Crime Reporting (UCR) Program Data: Arrests by Age, Sex, and Race, 1974-2021. https://www.openicpsr.org/openicpsr/project/102263/version/V15/view

Data

We used the following two variables at the county level.

Variables* Year Unit
Rate of Arrests for Violent Crime per 100,000 Black Population 1990 | 2000 | 2010 | 2020 Number
Rate of Arrests for Violent Crime per 100,000 White Population 1990 | 2000 | 2010 | 2020 Number

* Individuals from both Hispanic and non-Hispanic ethnicities are included.

Methodology

We calculated the Violent Crime Arrest Gap using a ratio formula:
$$
RPrBlWhViolent = \frac{PrBlViolent}{PrWhViolent}
$$

Where:
RPrBlWhViolent: Ratio of proportions of arrests of Black to White individuals
PrBlViolent = BlViolent / BlPop: Proportion of arrests of Black individuals for violent crime (BlViolent) to the total Black population (BlPop)
PrWhViolent = WhViolent / WhPop : Proportion of arrests of White individuals for violent crime (WhViolent) to the total White population (WhPop)

Data on four violent crime offenses, murder, rape, robbery, and aggravated assault for Black and White populations were filtered from the yearly crime data reported by agencies. The counts for these four offenses were added to get the count for violent crimes by each race (Black and White). The yearly summarized data does not have the county FIPS codes for all agencies. So, the filtered dataset was joined to the monthly dataset (https://www.openicpsr.org/openicpsr/project/102263/version/V15/view) to get the county FIPS code for all agencies using the originating agency identifier code. The count for violent crimes by race was aggregated by county using the county FIPS code.

Missing Data

The missing data were imputed in the following steps.

  1. For 2020, the states of Illinois and Florida did not have any data.
  2. We used the agency-level yearly summarized violent crime arrest data from ICPSR. To minimize missing or NULL values, we used the county-level aggregated (all races combined) data to fill the counties with a 0 count of arrests in our dataset for the years 2010, 2000, and 1990. The logic behind this is that if the total count of arrests for all races is 0, the count of arrests for Black and White populations should also be 0. Using this logic, 238 counties in 2010, 303 in 2000, and 338 in 1990 were filled with 0 in place of NULL values. We did not find the county-level aggregated crime data for 2020.
  3. After the imputation from the county-level datasets, the remaining missing values were filled using the median value of the 8 nearest neighbors in the same state. The eight nearest neighbors were identified using the KDTree algorithm in Python’s scipy.spatial module 7.
  4. After imputing missing data form steps 1-3, we have 260 counties (Illinois, Florida, and Islands) with no data in 2020, 78 counties (Islands) with no data in 2010, 24 counties (Kentucky) with no data in 2000, and 22 counties (Georgia) with no data in 1990.

Limitations

A large amount of missing data in the year 2020, 260 counties, including all the counties in the states of Illinois and Florida.

References

1. Federal Bureau of Investigation. (2019). Crime in the United States, 2019: Property crime. U.S. Department of Justice.

2.Beck, A. J. (2021, January). Race and ethnicity of violent crime offenders and arrestees, 2018 (NCJ 255969). U.S. Department of Justice, Bureau of Justice Statistics. https://www.bjs.gov/content/pub/pdf/revcoa18.pdf

3. Ulmer, J. T., Harris, C. T., & Steffensmeier, D. (2012). Racial and ethnic disparities in structural disadvantage and crime: White, Black, and Hispanic comparisons. Social science quarterly, 93(3), 799-819.

4. Krivo, L. J., Peterson, R. D., & Kuhl, D. C. (2009). Segregation, racial structure, and neighborhood violent crime. American journal of Sociology, 114(6), 1765-1802.

5. Henderson, H., Bourgeois, J. W., Smith, S., Ferguson, C. J., & Barthelemy, J. (2024). Police shootings, violent crime, race and socio‐economic factors in municipalities in the United States of America. Criminal Behaviour and Mental Health, 34(3), 296–310. https://doi.org/10.1002/cbm.2333

6. Manson, S., Schroeder, J., Van Riper, D., Knowles, K., Kugler, T., Roberts, F., & Ruggles, S. (2024). IPUMS National Historical Geographic Information System: Version 19.0 [Dataset]. IPUMS.

7. Virtanen, P., Gommers, R., Burovski, E., Oliphant, T. E., Weckesser, W., Cournapeau, D., … & Feng, Y. (2021). scipy/scipy: SciPy 1.6. 0. Zenodo.