According to the analysis of the 2009 raw data by the Center for Constitutional Rights, nearly 490,000 blacks and Latinos were stopped by the police on the streets last year, compared with 53,000 whites.
But once stopped, the arrest rates were virtually the same. Whites were arrested in slightly more than 6 percent of the stops, blacks in slightly fewer than 6 percent.
A check on whether or not there is discrimination is if the marginal Terry stop yields the same arrest rate. If marginally frisked whites are arrested at two times the rate than marginally frisked blacks, then this is an indication that there is statistical discrimination against blacks (not whites!)--it would be more efficient to stop more whites. But if marginal stops in the two groups yielded the same arrest rates, then we are efficiently discriminating.
While we cannot observe margins, the fact that we are gaining the same proportion of arrests post-stop is encouraging--if the average stop is approximately the marginal, then the article gives evidence that police in New York are discriminating efficiently.