Locking Up

Kevin Lewis

February 12, 2021

Who Punishes More? Partisanship, Punitive Policies, and the Puzzle of Democratic Governors
Anna Gunderson
Political Research Quarterly, forthcoming


The growth of the carceral state over the last few decades has been remarkable, with millions of Americans in prison, jail, on parole or probation. Political science explanations of this phenomenon identify partisanship as a key explanatory variable in the adoption of punitive policies; by this theory, Republicans are the driving force behind growing incarceration. This article argues this explanation is incomplete and instead emphasizes the bipartisan coalition that constructed the carceral state. I argue Democratic governors are incentivized to pursue more punitive policies to compete with Republicans when those Democrats are electorally vulnerable. I test this proposition using a series of regression discontinuity designs and find causal evidence for Democrats’ complicity in the expansion of the carceral state. Democratic governors who barely win their elections outspend and outincarcerate their Republican counterparts. This article highlights Democrats’ role as key architects in the creation of vast criminal justice institutions in the states when those Democrats are electorally vulnerable.

Violent Crime and the Overmilitarization of US Policing
Federico Masera
Journal of Law, Economics, and Organization, forthcoming


Using new data at the police department level, I propose an identification strategy for estimating the causal effect that police militarization has on reducing violent crime. I show that previous estimates are likely to be contaminated by unobserved factors that simultaneously determine militarization and violent crime. Upon addressing this issue, I find a point estimate that is 20 times larger than those estimated previously. I then find that one-fourth of the effect of militarization is due to the displacement of violent crime to neighboring areas. Police departments overmilitarize because they do not consider this externality. These new findings have significant implications for the policy debate concerning the costs and benefits of police militarization

Racist Cops, Vested “Blue” Interests, or Both? Evidence from Four Decades of the General Social Survey
Vincent Roscigno & Kayla Preito-Hodge
Socius: Sociological Research for a Dynamic World, January 2021


The murder of George Floyd by a Minneapolis police officer triggered U.S. and worldwide protests - protests that raised questions about police funding, use of force, and whether police officers are distinctly racist. In this article, the authors draw on nearly four decades of the General Social Survey to examine trends over time and specifically model whether those in law enforcement are more likely to hold racialized and arguably racist views, vested “blue” occupational interests, or both. Trends show declining public support for police expenditure and police use of force over time. The authors’ further modeling highlights stark differences between police and the general public, as well as between cops and those of similar occupational status. Specifically, police uniquely believe that they should receive more funding and have the right to use physical force against citizens; they are also more racist, a pattern especially apparent among white male officers. These findings, which largely support the arguments of current Black Lives Matter protesters, show how vested occupational interests and racialized orientations intersect in important ways, sometimes with perilous consequences.

Changes in Shooting Incidence in Philadelphia, Pennsylvania, Between March and November 2020
Jessica Beard et al.
Journal of the American Medical Association, forthcoming

Methods: This study accessed data from the Philadelphia Police Department’s registry of shooting victims from January 1, 2016, through November 26, 2020. This registry is updated daily and includes all individuals shot and/or killed with a firearm as a result of interpersonal violence. There were no changes in data collection policies or practices in 2020. Compared with trauma center records, the police registry contains approximately twice the number of individuals shot with a firearm. Changes in counts of individuals shot per week were examined following 3 time points: the enactment of Philadelphia’s first COVID-19 containment policy (closure of nonessential businesses; March 16, 2020), the killing of George Floyd (May 25, 2020), and the partial lifting of containment policies (June 26, 2020). A time-series analysis was conducted using autoregressive integrated moving average (ARIMA) models. The dependent measure, a count of individuals shot per week, was log-transformed to ensure normal distribution. Interruptions were specified the weeks of March 16, May 25, and June 26, 2020. We identified the ARIMA model by systematically testing local and 52-week seasonally lagged autoregressive, difference, and moving average terms to remove trends and account for temporal dependencies. Three transfer functions assessed associations between the interruption variables and the dependent measure, capturing abrupt permanent relationships (stepped change), gradual permanent relationships (asymptotic growth curve), and abrupt temporary relationships (immediate increase followed by gradual decline). We selected the best fitting model based on ARIMA terms and transfer functions with 2-sided P values of less than .05, autocorrelation and partial autocorrelation functions within 95% CI bands (calculated using the Bartlett formula), and low values for both the Ljung-Box Q statistic at 24 lags and the Akaike information criterion. Statistical models were estimated using SCA Workbench, version 6.3 (Scientific Computing Associates Corp). The Temple University institutional review board determined this study was not human subjects research and did not require approval.

Results: During the 256 weeks included in the study, 7159 individuals were shot in Philadelphia, which breaks down to a mean (SD) of 24.9 (7.1) individuals shot per week during the 219 weeks before enactment of COVID-19 containment policies and 46.4 (13.5) individuals shot per week in the 37 weeks after containment policy enactment. The ARIMA model that best fit this time series included local autoregressive and moving average terms and a gradual permanent association beginning the week COVID-19 containment policies were enacted. The killing of George Floyd and the partial lifting of containment policies were not independently associated with any changes in shooting incidence.

Intensified Scrutiny and Bureaucratic Effort: Evidence from Policing and Crime After High-Profile, Officer-Involved Fatalities
Deepak Premkumar
Public Policy Institute of California Working Paper, December 2020


This paper provides the first estimates of how high-profile, officer-involved fatalities (OFs) affect the arresting patterns for the involved police department and crime in that jurisdiction. To address the simultaneous effects that could occur after an OF -- (1) greater scrutiny of police, (2) reduced community cooperation with identifying and locating suspects, (3) reduced civilian crime reporting, and (4) changes in offending behavior -- I develop a theoretical model of policing behavior to provide empirical predictions of the changes in arrests due to each of the four possible channels. Following a high-profile OF, theft arrests drop by 3-11%, while arrests for the least serious offenses (e.g., marijuana possession and disorderly conduct) see sharp declines of up to 33%. Notably, arrests do not change for violent crime or more serious property crimes. These findings are consistent with scrutiny as the causal channel for the reduction in arrests. While the decline in arrests for theft is temporary, it persists for the least serious offenses, representing a sustained transition to a lower equilibrium effort. I also find substantial increases in offending: There is a significant rise of 10-16% in both murders and robberies. There are also smaller increases of 3-7% in theft and motor vehicle theft.

Masculinity, Ritual, and Racialized Status Threat: Examining Mass Shooter Manifestos Using Structural Topic Models
Jessica Pfaffendorf, Andrew Davis & Alexander Kinney
Sociological Inquiry, forthcoming


Both popular media and research often frame mass shootings as an individual issue having to do with mental illness or other individual differences. This work has unfolded in much the same fashion as that on other negative or anti‐social behaviors - such as the individual pathologization of suicide or rape. However, what this work has shown empirically is that there are often a set of circumstances that are uniquely social that motivate such actions. Following work in sociology, which offers social psychological and cultural explanations for gun violence, we argue that mass shooter motivations reflect social conditions -- especially those that instantiate toxic masculinity, social exclusion, and racism -- conducive to these events. This article uses a computational textual modeling approach to analyze the distinct social logics that motivate mass shooters. To do this, we identify a sample of 27 publicly available mass shooter “manifestos,” or documents left behind by shooters following their actions. Using topic models, we show that mass shooters exhibit a variety of preoccupations that underlie their actions. While shooters can exhibit a multitude of possible motivations, we find that expressions of masculine overcompensation, ritualistic responses to exclusion, and racialized status threat are prominent features of mass shooter manifestos, corroborating recent sociological explanations of mass shootings.

Did de‐escalation successfully reduce serious use of force in Camden County, New Jersey? A synthetic control analysis of force outcomes
Li Sian Goh
Criminology & Public Policy, forthcoming


Despite the widespread interest that de‐escalation training has attracted in law enforcement contexts over the past few years, we know little about its effectiveness in reducing use of force incidents. This study seeks to ascertain the effect of de‐escalation training on serious use of force events in Camden, a high‐crime and high‐poverty city in New Jersey. An analysis of individual officers suggested de‐escalation training had no significant effects on serious force, whereas a synthetic control analysis of the entire department suggested that de‐escalation training led to a 40% reduction in serious force events. It is suggested that spillover effects between trained and untrained officers may account for the discrepancy.

Airbnb and Neighborhood Crime: The Incursion of Tourists or the Erosion of Local Social Dynamics?
Laiyang Ke, Dan O'Brien & Babak Heydari
Northeastern University Working Paper, November 2020


The proliferation of internet-based home-sharing platforms like Airbnb has raised heated debates, with many in the general public believing that the presence of Airbnb can lead to an increase in crime and disorder in residential neighborhoods. Despite the importance of this debate to residents, policymakers, and other stakeholders, few studies have examined the causal linkage between Airbnb and crime in neighborhoods. We conduct the first such empirical test in Boston neighborhoods, focusing on two potential mechanisms: (1) the inflow of tourists might generate or attract crime; and (2) the creation of transient properties undermines local social dynamics. Corresponding to these mechanisms, we examine whether the number of tourists (approximated with reviews) or the prevalence of listings predict more incidents of private conflict, social disorder, and violence both concurrently and in the following year. We find evidence that increases in Airbnb listings--but not reviews--led to more violence in neighborhoods in later years. This supports the notion that the prevalence of Airbnb listings erodes the natural ability of a neighborhood to prevent crime, but does not support the interpretation that elevated numbers of tourists bring crime with them.

The online behaviors of Islamic state terrorists in the United States
Joe Whittaker
Criminology & Public Policy, forthcoming


This study offers an empirical insight into terrorists’ use of the Internet. Although criminology has previously been quiet on this topic, behavior‐based studies can aid in understanding the interactions between terrorists and their environments. Using a database of 231 US‐based Islamic State terrorists, four important findings are offered: (1) This cohort utilized the Internet heavily for the purposes of both networking with co‐ideologues and learning about their intended activity. (2) There is little reason to believe that these online interactions are replacing offline ones, as has previously been suggested. Rather, terrorists tend to operate in both domains. (3) Online activity seems to be similar across the sample, regardless of the number of co‐offenders or the sophistication of attack. (4) There is reason to believe that using the Internet may be an impediment to terrorists’ success.

A National Study of Child Maltreatment Reporting at the County Level: Interactions Among Race/Ethnicity, Rurality and Poverty
Brenda Smith et al.
Children and Youth Services Review, forthcoming


Child maltreatment report rates vary widely among states and counties. A portion of the variation likely reflects varying community-level risk and protective factors, but the variation also likely reflects community characteristics unrelated to risk, raising questions of equity and justice, in addition to safety. This study builds on previous research that focused only on the U.S. South to investigate nationally how county racial/ethnic compositions, poverty rates, and rurality relate to child maltreatment report rates. Aggregated county-level data from the National Child Abuse and Neglect Data System (NCANDS) from nearly all U.S. counties (n = 2,966 in 2015) were linked to data from the U.S. census and other sources. Bivariate tests and multi-level regression models assessed county-level factors associated with the child maltreatment investigated report rate. Consistent with previous studies, despite higher child poverty rates, child maltreatment report rates were lower in rural counties with majority populations of color compared to other counties. In addition, although county-level child poverty rates were generally positively associated with child maltreatment report rates, child poverty was not positively associated with child maltreatment report rates in rural counties with majority populations of color, primarily due to low report rates in rural counties with majority Black populations. To our knowledge, this is the first national child maltreatment study to disentangle county rurality from racial/ethnic composition by specifically investigating rural counties with majority populations of color.

Digitizing and Disclosing Personal Data: The Proliferation of State Criminal Records on the Internet
Sarah Lageson, Elizabeth Webster & Juan Sandoval
Law & Social Inquiry, forthcoming


Digitization and the release of public records on the Internet have expanded the reach and uses of criminal record data in the United States. This study analyzes the types and volume of personally identifiable data released on the Internet via two hundred public governmental websites for law enforcement, criminal courts, corrections, and criminal record repositories in each state. We find that public disclosures often include information valuable to the personal data economy, including the full name, birthdate, home address, and physical characteristics of arrestees, detainees, and defendants. Using administrative data, we also estimate the volume of data disclosed online. Our findings highlight the mass dissemination of pre-conviction data: every year, over ten million arrests, 4.5 million mug shots, and 14.7 million criminal court proceedings are digitally released at no cost. Post-conviction, approximately 6.5 million current and former prisoners and 12.5 million people with a felony conviction have a record on the Internet. While justified through public records laws, such broad disclosures reveal an imbalance between the “transparency” of data releases that facilitate monitoring of state action and those that facilitate monitoring individual people. The results show how the criminal legal system increasingly distributes Internet privacy violations and community surveillance as part of contemporary punishment.

Targeting Intensive Job Assistance to Ex-Offenders by the Nature of Offense: Results from a Randomized Control Trial
Christopher Bollinger & Aaron Yelowitz
University of Kentucky Working Paper, January 2021


As many as two-thirds of newly-released inmates will be arrested for a new offense within three years. This study evaluates the impact of job assistance on recidivism rates among ex-offenders. The job assistance program, run though the private company America Works, uses a network of employers to place clients. Ex-offenders were randomly assigned to intensive job assistance (treatment group) or the standard program (control group). The intensive program is meant to improve average work readiness for ex-offenders. It reduces the likelihood of subsequent arrest among non-violent ex-offenders, but has little effect on violent ex-offenders. The re-arrest rate for non-violent ex-offenders in the treatment group was 19 percentage points lower than those in the control group. The re-arrest rate for violent ex-offenders in the treatment group was indistinguishable from those in the control group. We estimate benefits from intensive job assistance from averted crimes and find that they outweigh the $5,000 up-front cost for non-violent ex-offenders.


from the


A weekly newsletter with free essays from past issues of National Affairs and The Public Interest that shed light on the week's pressing issues.


to your National Affairs subscriber account.

Already a subscriber? Activate your account.


Unlimited access to intelligent essays on the nation’s affairs.

Subscribe to National Affairs.