Randomizing at the police station level isolates trained officers from comparison personnel.

Download image · 1200 × 630 PNG

Research allocation

Randomly assign police stations

419 police stations in Bihar

217 training stations
Receive the training intervention.
202 comparison stations
Form the comparison group.

Illustrative summary of the source design: Allocation is at station level. The source reports ongoing research; no outcomes are shown.

See source →

How

When to use

In law enforcement evaluations, individual officer randomization risks contamination when officers interact at the same station. Station-level cluster allocation prevents spillovers between trained and untrained personnel. Use randomized cluster allocation across police stations to rigorously evaluate institutional training programs.

Randomly assign police stations to intervention or comparison groups to evaluate training impacts without spillover.

  1. Map and compile a roster of participating police stations across the target region.
  2. Set cluster boundaries at the police station level to contain daily interactions within stations.
  3. Assign unique identification codes to all listed police stations.
  4. Generate a random allocation sequence using statistical software.
  5. Assign 217 police stations to the training intervention group and 202 stations to the comparison group.

Why it works

This cluster allocation prevents information spillovers and allows researchers to measure true organizational shifts.

Example

Researchers randomly assigned 217 police stations to the intervention group and 202 police stations to the comparison group in Bihar, India.

About the study

Participants

3,526 police officers in 419 police stations across Bihar, India.

Study Design

Cluster randomized controlled trial allocating 419 police stations into treatment and comparison groups.

Results

Research ongoing; results forthcoming.

Source

See source →