The great disappearance of violent crimes in Illinois

The FBI is Crime data explorer can be difficult to useBut there can be good data if you know where you have to search. One of the characteristics I like is that Data acquisition tools This allows you to see the reported crime on an agency, a state or local level in 1986.

However, this for Illinois crimes is also revealing about the possible pitfalls of our crime data report system. See, the state of Illinois' violent crime trend since 2010:

This is a great decline in 2021, followed by a considerable increase of 2022 and 2023, albeit at much lower levels than before the waste. But US crime crimes in Illinois now really well below 2020 level?

Well, no. Unfortunately, this secret has a relatively simple answer, which can be summarized as: Chicago.

One of the challenges in cooperation with the crime data explorer is that you do not say when data is missing. Chicago therefore reports only 6 months of data in 2021 due to the NIBRS transition in the CDE data acquisition tool. CDE offers the option of looking for monthly data and the lack of data between January and June 2021 in Chicago is obvious there.

The drop in 2021 is therefore easy to explain as a NIBRS transition quirch. But what about the lower level in 2022 and beyond?

This seems to be a change in the way Chicago counted heavy attacks.

If you look chicago's crimes in the view Real-time criminal index You see no obvious changes in the counted crime in Chicago. Here you will find heavy attacks in Chicago that roll over 12 months, which makes a large proportion of Illinois crimes. There is no enormous decline, in fact, stricter attacks in Chicago rose somewhat compared to the pre-Kovid values ​​after the covise.

Here you will find the monthly crime counts of serious physical injury in Chicago from 2018 to 2023 in the CDE.

There is no data for the first 6 months of 2021, followed by an enormous decline in severe attacks. The difference between the CDE and RTCI Illinois State Police While the previous ones come from the Chicago police Open data portal.

I started writing this piece a few weeks ago and dealt with data from Illinois State Police to show the differences in the monthly tightened attacks between the FBI, Illinois State Police and the CPD open data. There is no data from the state police before July 2021, but it can be clearly seen that severe attacks in Chicago remained relatively stable in the open data of the city from 2018 to 2024. The enormous decline in the number of serious attacks in Chicago was a function of far fewer crimes reported by the city of the state police and then the FBI after the NIBRs had been implemented.

Here I would normally provide the final thoughts about the challenges of working with crime data and the importance for the analysis of violent criminal trends in Illinois. National crime estimates are exactly that and we should take on uncertainty if we discuss the changing national trends. I would also find that Chicago is not alone in this problem and point out how “Human mistake” In Oakland, the US power of power crime rate rose by 1 percent in 2023.

All of this is good and good, but a strange thing happened when I sat down last week to end this piece: Chicago fixed the problem.

While the number of serious attacks in Chicago was around 300 per month in the Illinois State Police data a few weeks ago, it is now well over 1,000 per month. So much so that the revised numbers agree with what is reported by CPD (sorry for all colors in this diagram!):

The figures reported by the ISP from 2021 are still not good, but in 2021 there was a catastrophic year for reporting on crime data everywhere. The 2022 to present data looks exactly on the right point.

So what's next?

Chicago's figures should be revised in the next publication of the FBI in the next publication of the uniform criminal report, even though this process is opaque, so it is difficult to say. The under -reporting in Chicago and the transfer to Oakland should be canceled more or less (if not entirely), so I would not expect violent crimes at the national level. But on the other hand, you could be repaired or not, or there could be another problem that has not yet come to light.

In any case, this is a good example of the living nature of the crime data. There are enough mistakes and enough uncertainty that it is never complete, even if it is complete. And finally it is a good lesson in the search for data reporting problems if you find a reported count that is too good (or bad) to be true.

Leave a comment