Stripping Away Fictions in the Independent Monitor’s Report on Policing in Chicago

Flawed report seeks to advance the canard of racist policing in Chicago

Back in June, the independent monitor appointed to oversee implementation of the multi-year consent decree for the Chicago Police Department (CPD), Maggie Hickey, released her "Consultant Report" which makes several hyperbolic claims over CPD stops. Chicago's politicians wasted no time endlessly and breathlessly repeating those claims over and over and with the cooperation of an obliging media. Thus far, it does not appear that anyone has bothered to look into the veracity of the allegations. Such is the sorry state of establishment media in this town.

Verifying, or refuting, the report's claims is actually not difficult. With a slightly dated computer, a rudimentary understanding of basic arithmetic, and a little bit of time, an analysis can be performed by virtually anyone. The Chicago Police Department makes the Investigatory Stop Report (ISR) data — the basis for the report's claims — publicly available on its website. An ISR occurs when an officer questions an individual suspected of involvement in a crime or who police suspect is about to carry out a crime. Given the number of records, the files are not small; but they're not unwieldy large either.

In the CPD files, the data is arranged in a standard format. There are 169 columns, each of which is a field on the ISR form. The fields contain information about the time and location of the stop, why it occurred, who was involved, if any contraband was found, and how the stop concluded. There is one record for each individual who was stopped. Each record has its own row in the file.

Taking a cursory glance at the CPD files from 2018, 2019, and 2020, it is impossible to confirm what data was used in the Consultant Report, which states records were examined over the three-year period. However, 128,728 records were stated to have been reviewed from 2018, yet the current 2018 file from CPD contains a total of 132,557 records. The other years documented in the report are similarly off. It is unlikely a few thousand records would significantly throw off the analysis. Nonetheless, it does make one wonder why the numbers don't match.

In working with the 2018 file, there are 91,554 ISR records for blacks, 11,446 records for whites, and 28,034 records for Latinos. The balance of records are categorized with other races and ethnicities. This ratio, of blacks to whites to Latinos, is the basis for nearly all the erroneous claims to be found in the report. Because blacks are overrepresented when compared to the overall population, CPD must be racist. At least that is the claim -- one entirely without merit.

There are numerous problems with this assertion, not the least of which is that it's statistically unsound. Because ISR records are not a random sampling, there is no reason to expect the race ratio to match the overall population. Consider two locations: Restaurant row on Randolph Street, and the United Center during a Bulls game. Where would you expect to find more red jerseys? ISR records are like counting red jerseys during a Bulls game.

Further, the report appears not to understand the statistical nature of sets and subsets. The repeated claims of racism are based on the anticipation that subsets of a population should match the general population. This is simply untrue. People attending a Bulls game are a unique subset of the general population. Those wearing red jerseys at the Bulls game are part of a further subset. Within that group you will find some portion holding a beer. And again, some fraction of those will be drunk.

All whales are mammals, but not all mammals are whales. If you sample a group of people at a Bulls game, wearing a red jersey, holding a beer and determine that five percent of them are drunk, that does not mean that five percent of the general population is drunk. It also does not mean that five percent of people at the Bulls game are drunk. Even if they are wearing a red jersey.

Wherever the report makes a claim that some subset of data should match the overall population, that claim should immediately be disregarded as unequivocally false.

Digging into the file for the year 2018, the data provides the reason for the stop. The breakdown of reasons as a percentage of stops for each race is in the following table:
‍

‍

There is no indication of racism here. Police stop and interview blacks, whites and Latinos in the relatively same proportion for all reasons. In fact, the numbers match so closely it would be impossible to make them up. Even if every police officer was lying, it would be statistically impossible to get the numbers this close given the fact that there are thousands of officers filling out over 100,000 contact cards in 2018 alone. Any suggestion that officers stop and interview blacks based on their race is simply not supported by facts.

The data also allows analysis on pat downs and searches. The Consultant Report makes several claims not borne out by the data. As a result of pat downs in 2018, a firearm was recovered from blacks 2.33 percent (720 of 30,887) of the time, for whites the rate was 1.62 percent (30/1,852), and for Latinos 1.63 percent (136/8,330).

In the following table are the results of contraband recovered as a result of pat downs for each race:

‍

Whereas a search is more intrusive than a pat down, the hit rate on searches is a little higher. For blacks, a firearm was recovered 5.52 percent (986/17,847) of the time. For whites and Latinos, the firearm recovery rate was 2.71 percent (40/1,477) and 4.97 percent (211/4,242) respectively.

This table shows contraband recovery by race as a result of a search:

‍

The ISR data also contains information on whether any enforcement action was taken. The enforcement rate for blacks was 26.68 percent (24,431/91,554). It was 22.03 percent (2,522/11,446) for whites, and 24.51 percent (6,870/28,034) for Latinos.

Moreover it's important to point out that 2018 is not unusual or an outlier. The tables below show the consistency of the data through 2019 and 2020.

The bottom line is that the Consultant Report makes several claims that are not supported by facts.

‍