Guessing Covid may be a factor but also might just be 2020 and early 2021

Your sources say there have been 22 shootings this year but that is not what your data shows. https://www.edweek.org/leadership/school-shootings-this-year-how-many-and-where/2024/01. What’s up with that?

@Rolledupaces A school shooting includes any incident where there was gunshots at a school. The majority of these are personal beefs or gang violence, but things like negligent discharges, suicides, or stray bullets from off campus can also be included. An active shooter incidents is one where the shooter is attempting to kill people in a confined space, usually without a pattern or method to their selection of victims.

Statista lets you separate school shootings from active shootings, but their data stops halfway through 2022. Ed Week lists all the shooting incidents, but doesn't clarify if they were active shooter incidents. So I looked at the reports and googled the incidents with the most fatalities to see if they were active shootings. Only one incident, Perry High School, involves a shooter randomly killing without regard to victim selection. The other incident seem to be personal beefs or gang related, and the Perry incident was the only one in which multiple people victims were killed.
Still, this is a less rigorous way to count the number of active shooters. I wish I had another source that consistently counted active shooting incidents at school by year.

To the people voting for "No trend, just noise", do you expect 2025 to have 4 or more active shooting incidents, as 57% of years 2004-2024 have?

@GG How about doing some actual statistics instead of drawing a line where you think it looks nice? To my eye, this looks right out of a Poisson-distribution, with a by-eye p=0.1. Would be fun to see if this guess is actually correct lol.

@sbares I asked ChatGPT to run a linear regression on years 2018-2024, with 2020 excluded due to the schools being shut down half the year. It found a slope of -1.6, with a correlation of -0.926, and a p-value of 0.0081

Here's the data so you can run your own analysis:

2024: 1, 2023: 1, 2022: 3, 2021: 8, 2020: 1, 2019: 7, 2018: 11, 2017: 6, 2016: 4, 2015: 4, 2014: 4, 2013: 4, 2012: 5, 2011: 2, 2010: 7, 2009: 1, 2008: 1, 2007: 4, 2006: 8, 2005: 3, 2004: 2

@GG I.e. you asked it to draw a line where you thought it looked nice. I think my eyeballing is as good as yours, but I can't be bothered to do the analysis to back it up, that would take several minutes of my life.