DATA THREAD The Massachusetts Trial Court has statewide charge level data from the courts, for FY 2019 & 2020. There are 312,087 cases in this dataset, each with race, age at arraignment, lead charge, sex, and degree of offense. Follow this thread for a brief analysis 1/x

2/x Firstly, the data is available here: https://www.mass.gov/info-details/trial-court-statistical-reports-and-dashboards

3/x Here are the number of cases present, by Court division.

4/x Now, I mentioned there's demographic information, as well as charging information. I'll start with the latter.

5/x The data *only* contains the lead charge for each case, which is a bummer as we can't replicate the phenomenal work by @HLS_CJPP to show how charging decisions by Prosecutors drives sentencing disparity. The Trial Court *does* keep this information. It has not been released.

6/x We can still learn a lot. For example, an overwhelming majority of cases in the Trial Court system are for MV offenses.

The percent of totals are:
44.05% - Motor Vehicle
19.69% - Person
14.87% - Property
10.11% - Other
03.11% - Public Order
01.12% - Weapons Offenses

7/x And so, here are some of the most common charges:

8/x No, we sadly have no dispositional information. So we are not able to statistically confirm the outcomes of these types of cases, and how those dispositions might vary from court to court, or county to county. This is a shame.

9/x *BUT* we have race and sex. So let's look at the same lead charges by those exact characteristics.

10/x Here, you'll find the number of cases by the lead charge, and the race of the defendant.

11/x But of course, that does not account for the massive population disparity between people of color and white people, in Massachusetts. Let's correct that.

12/x This accounts for the population. It may seem confusing! But basically read the per-100k as "Black people make up about 7% of the MA pop., but over 40% of all Motor Vehicle cases per 100,000 people in MA."

13/x Let's repeat the above, but for all cases collectively, broken down by the race & sex of the defendant where it's known.

Black men have the most number of cases, and white men the least, per 100,000 people.

14/x One of the things this dataset has is a unique defendant identifier, which for the first time should allow the public to visualize people who may have had multiple cases across the Commonwealth in the two-year window.

15/x There are 234,436 unique defendants in this dataset. About 48,227 people, or 20.5%, had more than one case. 2,052 (<1%) people had more than 10 cases.

16/x This data has information about the arresting PD. Soooo let's hypothetically assume you wanted to see which police departments are arresting people on principle drug charges in the year of our lord 2019....

17/x We also have the defendant's age at filing/arraignment! So let's dig deeper. It's widely known that warmer weather leads to more police incidents. Let's see how that filters down to the cases that end up being filed.

18/x Drug cases. The darker the color (i.e. the more orange), the more cases that were filed.

n = 22,028

19/x Motor vehicle cases

n = 137,476

20/x "Person" cases, or those with violence on another person.

n = 61,448

21/x "Property" cases, or those with destruction of property.

n = 46,415

22/x Public order cases.

n = 9,699

23/x Weapons cases, almost exclusively for possession under MGL c.269. Includes §10(a) & §10(f).

n = 3,489

24/x Honestly I am glad the Trial Court (finally) publicly released data. But it is for too small a window, and with too much missing. Yet, I hope this thread is useful to you in thinking about how you as a researcher, journalist or curious person can approach this data.

25/25 I also hope that you're able to understand how important this data is, after all: modern American policing was born out of the courts, not the police. How we rethink policing should be coupled with an equal movement for court reform. You can't manage what you don't measure.

Addendum #1: For the race category "white," I used the population of white people who are NOT of hispanic origin, from the latest 5-year ACS estimates.

Addendum #2: If you'd like to discuss this further send me an email nassers [at] protonmail [dot] com