Searches: the outcome test
The strongest test this data supports asks two questions at once: who gets searched, and whether those searches find anything. Searches of Black and non-Black stopped people find contraband at statistically indistinguishable rates. What differs is who gets searched at all: Black stopped people were pat-down searched at up to 4.54× the rate of non-Black stopped people, and above their rate in every district.
4.54×
The largest district pat-down ratio: Black pat-downs per stop divided by non-Black pat-downs per stop, within the same district.
Each district shaded by its pat-down rate ratio: darker means Black stopped people are searched at a higher multiple of the non-Black rate. Click a district for its full page.
Districts are listed in numeric order. This is not a ranking: each ratio compares the two groups within the same district.
| District | Black pat-down rate | Non-Black rate | Ratio | Black pat-downs (N) |
|---|---|---|---|---|
| District 1 | 7.0% | 2.1% | 3.28× | 1,600 |
| District 2 | 3.5% | 1.5% | 2.33× | 485 |
| District 3 | 4.8% | 1.7% | 2.79× | 1,124 |
| District 4 | 4.9% | 3.0% | 1.65× | 848 |
| District 5 | 5.7% | 1.2% | 4.54× | 1,800 |
| District 6 | 8.8% | 3.0% | 2.96× | 2,484 |
| District 7 | 10.2% | 5.2% | 1.94× | 2,888 |
Rates are per stopped person within each group, so a district's stop volume does not inflate them.
Searches: the outcome test
Where an outcome was recorded, searches of the two groups found something at rates that are statistically indistinguishable (13.9% vs 12.5%). The disparity is in who gets searched at all: see the per-district search rates. This is the “outcome test”: it needs no population baseline, because the denominator is the searches themselves.
| Search type | Black searched (N) | Black hit rate | Non-Black searched (N) | Non-Black hit rate |
|---|---|---|---|---|
| Consent searches | 1,034 | 11.0% | 94 | 11.7% |
| Warrant searches | 1,048 | 11.7% | 72 | 13.9% |
| Protective pat-downs | 11,324 | 14.3% | 915 | 12.5% |
| All (with recorded outcome) | 13,406 | 13.9% | 1,081 | 12.5% |
Hit rate = searches where something was seized ÷ searches with a recorded outcome. Probable-cause searches carry no outcome field in MPD's data, so they appear in search rates but cannot appear here. Searches with no recorded outcome (0) are excluded from the denominator and reported, never guessed. Citywide figures: search outcomes are too thin to split by district honestly.
Related: Stop representation by district | All analyses