Benchmarking Portland Police Bureau Traffic Stop and Search Data

Problem

The problem the study aimed to address: 

The study addresses the challenge of identifying racial or ethnic disparities in traffic stop and search data by the Portland Police Bureau (PPB). Specifically, it seeks to improve benchmarking methods to determine whether disparities indicate police bias or are influenced by external factors.

General impact on the system and/or public: 

Understanding and addressing disparities in police traffic stops can improve community trust, refine law enforcement practices, and ensure fair treatment across racial and ethnic groups.

Research Questions:

1. What are the most valid methods for analyzing racial/ethnic disparities in traffic stops and searches?
2. How can the PPB improve its data collection, analysis, and reporting systems to address potential bias more effectively?
    

Method and Analysis

Program Evaluated: 

The study evaluates the PPB's traffic stop data collection and analysis practices, focusing on limitations in existing benchmarking methods and exploring alternative strategies.
 

Data and Sample Size: 

• Traffic stops recorded by the PPB from January 2004 to June 2008, totaling 361,389 stops (81% of all recorded stops).
• Supplementary data sources included Portland Census demographics, violent crime data, and citizen-initiated police service calls.
   

Analysis Used:

• Evaluation of various benchmarking strategies, including census comparisons, daytime versus nighttime stops, traffic unit versus regular patrol stops, and multivariate logistic regression for search decision-making.
• Examination of geographic and demographic factors influencing exposure to law enforcement.
   

Outcome

Key Findings:

• Data Challenges: Significant portions of traffic stop data were incomplete or lacked key identifiers, limiting the analysis.
• Exposure Disparities: African American drivers were disproportionately stopped due to increased police activity in neighborhoods with higher crime rates and service calls, rather than systemic bias.
• Benchmarking Gaps: Traditional benchmarks like census data were insufficient for accurately assessing bias; alternative methods provided better contextual insights.
• Search Disparities: African Americans, Hispanics, and Native Americans faced higher rates of discretionary searches than Whites, although race alone was a weak predictor.
   

Implications or Recommendations: 

• Data Improvements: PPB should enhance data collection, including linking traffic stop data with CAD and citation systems and recording more detailed stop reasons and search types.
• Benchmarking Strategies: Employ multiple benchmarking methods to address disparities and rule out alternative explanations, such as differential exposure to law enforcement.
• Community Engagement: Maintain open communication with communities experiencing higher police presence to explain enforcement strategies and share successes in crime reduction.

This study underscores the importance of robust data and multifaceted analysis in addressing concerns of racial equity in law enforcement practices.