On Tuesday, the city released an independent analysis that concludes there is no evidence of racial “process bias” in how Cleveland police carry out stops, searches, and seizures. That finding lands in a city where other reviews, including a News 5/The Marshall Project data investigation and a federal consent decree monitor’s assessment, continue to show sharp racial gaps in who is stopped and searched.
In the city’s news release, Dr. Tanaya Devi, co‑founder and chief data scientist of Sigma Squared, said, “While the data reveal disparities in search rates, they do not show evidence of process bias in overall search decisions,” according to the City of Cleveland. The administration framed the Sigma Squared work as part of a push for near‑real‑time analytics to monitor stops and improve policing practices.
News 5 and Marshall Project findings
An independent review by News 5 Cleveland and The Marshall Project examined nearly 17,000 encounters and found that Black people made up almost 63% of stops and were searched at least three times more often than white individuals. When searches did occur, contraband was recovered about 37% of the time from Black subjects versus 32% for white subjects. Jeffrey A. Fagan, a Columbia Law School professor who reviewed the data for that investigation, said that pattern, more searches of Black residents without a higher hit rate, “suggests they’re exercising some kind of racial discrimination.”
What Sigma Squared tested
Sigma Squared’s full report applies a pipeline of controlled analyses and industry‑standard hit‑rate and threshold tests that account for location, stop reason, time, and other context. After adjusting for those factors, the firm found that Black subjects were still searched more frequently in comparable contexts, about 50.5% more often in traffic stops, but said the hit‑rate and threshold tests did not identify process bias and cautioned that the administrative data cannot always distinguish discretionary searches or capture in‑stop dynamics, according to Sigma Squared.
Federal monitor flagged disparities
The Cleveland Division of Police Consent Decree Monitoring Team reported that Black drivers were more than 3.7 times as likely as white drivers to be stopped in 2024 and said disparities persist across stops, searches and arrests. The monitoring team stopped short of attributing the gaps to deliberate bias but urged more analysis, targeted training, and stronger supervisory review to address the gaps, according to the Consent Decree Monitoring Team.
Those three tallies, the city‑hired Sigma Squared tests, the monitoring team’s raw‑number assessment, and the local investigative reporting, are all looking at the same data and arriving at very different headline takeaways. That technical split has sharpened a public debate over whether statistical “no bias” findings settle questions of fairness on the street, or whether deeper audits and stronger oversight are still needed.
Legal implications
Legal scholars and the monitoring team note that large disparities in stops and searches can raise constitutional concerns even when hit‑rate tests look neutral, because those tests do not capture everything that happens during an encounter. The monitoring team explicitly recommended that the city investigate officer‑level patterns and strengthen supervisory oversight to guard against potential constitutional violations, the Consent Decree Monitoring Team says.
What comes next
City officials said they will use Sigma Squared’s dashboard for near‑real‑time analysis, bring the findings into community engagement events, and continue sharing data with federal monitors, as reported by News 5 Cleveland and The Marshall Project. Advocates say independent audits of body‑worn camera footage, supervisory approvals, and stop narratives remain necessary to explain why the disparities persist.
For Clevelanders, the tension between statistical measures and raw disparities means the numbers are unlikely to be the last word. Monitors, outside analysts, and residents will be watching whether audits, training, and policy changes actually narrow the gaps. Both the city and the monitors say they will keep examining the data and the decisions officers make on the street to determine whether further corrective action is required.
