Policing policies and police behavior significantly influence political, social and economic landscapes in the United States. Currently, police interactions with the public are under intense scrutiny. I use an RCT to document the possible presence of bias in police departments.
Experimental design
I will use a randomized controlled trial to measure police bias in local police departments across the United States.
Design details
Randomization method
Randomization done in office by a computer.
Randomization unit
Randomly selected police departments
Randomly assigned treatment of race and gender
Randomly assigned week, day and time of intervention
Randomly modified treatment text
Clusters
2,135
Observations
2,135
Arms
2,135 Departments are selected randomly across the United States
Race (White, Black and Hispanic) is assigned equally across departments. E.g., ~2,135/3 departments receive White treatment.
Gender (female and male) is assigned equally across departments. E.g., ~2,135/2 departments receive female treatment.
Departments are assigned to 10 weeks roughly equally. E.g., ~2,135/10 departments are assigned to week 1.
Departments are assigned to 3 weekdays (MTW) roughly equally. E.g., ~2,135/3 departments are assigned to Monday.
Treatment text is assigned equally across departments. E.g., ~2,135/2 departments receive "friendly" treatment.
Primary outcomes
The primary outcome of interest for this project concerns whether and how police departments respond to inquiries about officer-complaint forms. In addition to gaining an improved understanding of general response behavior, I seek to document whether departments' behaviors systematically differ across the perceived identity of the individual asking for the officer-complaint form.
Cited by
No citation data available for this registration yet.