Do Peers in the Police Academy Influence Officers' Future Policing Behavior?

This study investigates whether peer composition in the Chicago Police Academy influences officers’ future policing behavior. Using a lottery-based random assignment of recruits to cohorts, the study analyzes arrest data, officer demographics, and assignment records. Findings reveal that higher shares of female and older peers significantly reduce low-level arrests by approximately 6-13% per 100 shifts, suggesting peer influence on policing preferences. Black and minority peers also have some effect, though less precisely estimated. The study concludes that peer effects shape policing behavior primarily through preference spillovers rather than assignment changes, with effects persisting up to four years post-academy.

Full Citation and Link to Article

Rivera, Roman, "Do Peers Matter in the Police Academy?"AMERICAN ECONOMIC JOURNAL: APPLIED ECONOMICS (FORTHCOMING).https://www.aeaweb.org/articles?id=10.1257/app.20220348&&from=f#:~:text=I%20find%20that%20higher%20shares,effects%20of%20gender%20and%20age

Extended Summary

Central Research Question

The study by Roman Rivera investigates whether the composition of peer groups in the police academy influences officers’ future policing behavior, particularly their arrest practices. The research is motivated by the broader policy discussion around increasing racial and gender diversity in law enforcement to reduce aggressive policing, which disproportionately impacts minority communities. While previous studies have examined the direct effects of officer demographics on policing behavior, this paper explores whether exposure to certain types of peers during academy training can have long-term behavioral spillovers. Specifically, the research asks: Does the presence of more diverse peers—defined by gender, age, and race—affect how officers police once they are in the field?

Previous Literature

Existing literature has established that officer demographics significantly influence policing behaviors. Studies have found that female and minority officers tend to police less aggressively compared to their white and male counterparts (Ba et al., 2021; Goncalves and Mello, 2021; Hoekstra and Sloan, 2022). This has led to policy proposals advocating for more diversity in law enforcement to address concerns about excessive policing, racial bias, and community relations.

However, while these studies focus on individual officer behavior, the potential spillover effects of diversity within training environments remain less explored. Peer effects are well-documented in education and workplace settings, where individuals tend to adopt behaviors and preferences of those around them (Sacerdote, 2011; Angrist, 2014). In law enforcement, this could mean that exposure to peers who police less aggressively might lead to broader changes in arrest practices. The paper builds on this by leveraging the unique context of the Chicago Police Academy, where recruits are randomly assigned to cohorts, allowing for an examination of peer effects in a controlled setting.

Data

The study utilizes a rich dataset on officers from the Chicago Police Department (CPD), covering recruits who entered the police academy between 2009 and 2016. The dataset includes detailed information on:

• Officer Demographics: Race, gender, and age at the time of entry into the academy.

• Arrest Records: Number and type of arrests made by officers after graduating from the academy, classified into serious crimes (e.g., violent felonies) and low-level offenses (e.g., drug possession, disorderly conduct).

• Assignment Data: Information on officer shifts, units, and locations, allowing for a distinction between behavioral effects and assignment-driven effects.

• Peer Group Composition: The demographic makeup of each recruit’s academy cohort, which was determined through a lottery system, ensuring random assignment.

This dataset allows the study to isolate peer effects by removing concerns about self-selection into academy groups, as the lottery ensures that officers do not choose their peers.

Methods

The study employs an econometric approach to measure the impact of peer composition on future policing outcomes. The core analysis relies on a regression framework, where an officer’s arrest behavior is modeled as a function of their peer group characteristics, controlling for individual demographics and time effects.

The key independent variables are:

• Percentage of Female Peers: The share of women in an officer’s academy cohort.

• Percentage of Older Peers: The proportion of peers aged 27 or older at the time of recruitment.

• Percentage of Minority Peers: The share of Black and non-Black minority recruits in the cohort.

The primary dependent variable is the number of arrests per 100 shifts, separated into low-level and serious arrests. The study also decomposes the impact into two components:

1. Assignment Effects: The extent to which cohort composition influences where officers are placed after graduation (e.g., district and shift assignments).

2. Behavioral Effects: The extent to which officers adopt different policing preferences based on their academy peer group.

To further validate the findings, the author examines the persistence of peer effects over time and explores potential mechanisms, such as socialization and preference spillovers.

Findings/Size Effects

The study finds strong evidence that peer composition in the academy significantly influences future policing behavior, particularly for low-level arrests.

1. Gender and Age Effects:

   • A 5-percentage-point (pp) increase in the share of female peers in an officer’s academy cohort reduces low-level arrests by approximately 6% per 100 shifts.

   • A 10pp increase in the share of older peers (above age 27) reduces low-level arrests by about 13% per 100 shifts.

   • These effects persist for at least four years post-academy, suggesting that early-career peer influence has lasting consequences.

2. Race Effects:

   • Higher shares of Black peers also reduce low-level arrests, but the effect size is smaller and statistically noisier than the gender and age effects.

   • While there is some evidence that white officers exposed to more Black and non-Black minority peers arrest fewer minorities, the results are less consistent than those for gender and age.

3. Serious Arrests:

   • There is a small but positive effect on serious arrests, suggesting that officers are not simply policing less overall but are instead shifting focus from minor infractions to more significant offenses.

4. Assignment vs. Behavioral Effects:

   • The findings indicate that the changes in arrest behavior are driven primarily by shifts in officers’ preferences rather than differences in their work assignments.

   • Peer effects do not significantly influence the types of districts or shifts officers are assigned to, meaning the differences in policing behavior arise within the same working conditions rather than from external factors.

     
     

Conclusion

The study provides compelling evidence that the composition of police academy cohorts influences officers’ long-term arrest behaviors. The findings suggest that increasing diversity in training environments can have spillover effects that lead to less aggressive policing of low-level offenses without reducing attention to serious crimes.

These results have important policy implications:

1. Recruitment Strategies:

   • Efforts to increase gender and age diversity in police departments should consider not just direct representation benefits but also the broader impact of peer effects.

   • Hiring more female and older officers can create a cultural shift that reduces over-policing of minor offenses.

2. Training Environments:

   • Given that early-career experiences shape long-term policing behaviors, police academies should foster diverse and integrated training cohorts.

   • Structured interactions among recruits of different demographics may enhance socialization effects and reinforce norms of less aggressive policing.

3. Beyond Individual Hiring:

   • While many policies focus on hiring minority officers to improve policing outcomes, this study highlights the importance of who recruits train with, not just who is hired.

   • Creating balanced peer groups within the academy can help reshape policing norms across entire departments.

Overall, this study contributes to the understanding of how diversity in law enforcement affects policing outcomes, demonstrating that peer effects in the academy have long-term, measurable impacts on officer behavior. These insights reinforce the idea that addressing aggressive policing requires structural and cultural changes, starting at the very beginning of an officer’s career.