Can’t edit the post above, but in addition to various other typos, I erroneously indicated that of the 47 suspicious person reports which "involved descriptions in which race or ethnicity was identified,” 33 of the suspicious persons were identified as white. This should read 34. More importantly, while I am not questioning his motives, I want to clarify why I believe @hebegebe’s calculations are mistaken and misleading.
Hebegebe used the Smith student body for his baseline population. But, as @Catcherinthetoast pointed out, the demographics of the surrounding community are much different than the demographics of the Smith student body. So which should be used, the students or the City? The Report provides both, but for whatever reason Hebegebe only focused on the students. This makes no sense to me. Employees at a small college generally know the students, and would be much more likely to find outsiders suspicious.
The Reports bear this out. Smith is an all girl’s school, yet in 2015 around 85% of the reported “suspicious persons”were male. In other words, definitely not students.
Obviously the calls were much more likely to involve non-students, and the broader demographic provides a much better point of comparison
Here is the demographic breakdown in the actual report:
If we assume “completely race-neutral reporting” (as did Hebegebe) but we use the city demographics (as set out in the report) as our baseline . . .
- What are the odds that it is purely random chance that only 34 white people were reported, instead of the predicted 42 white people?
- What are the odds that it is purely random chance that 7 black people people were reported, instead just the predicted one black person?
It’s been too many decades since my only statistics course, but if I figured out how to use the online calculators correctly then it seems the odds against this being a random occurrence are very, very long. Don’t quote me on this, but I get around 1900-to-1 for the former and around 25000-to-1 for the latter. @hebegebe or anyone, how’d I do on my math?
The numbers are even more damning if we look at the review of the police logs for 2017-2018.
For the period reviewed, race, color or ethnicity were referenced in 46 incidents. Of the 46 incidents, there were eight incidents where the people were described as black or African-American; 31 incidents where the people were described as white or Caucasian . . . There were two incidents where it was noted that the subject of the call was “dark- skinned” or with “darker skin" . . .
So, of the 46 incidents, 10 involved black or “dark-skinned” people, when the demographics of the area predict it should be one. I’ll let one of the math people run the numbers on this one.
All that said, I still don’t think we need all thesestatistics to know that these things happen to black people a heck of a lot more than they happen to white people. Nonetheless, while I am sure it must have been inadvertent, Hebegebe has presented his numbers and calculations to create the false impression that the data supports a conclusion that there is no disparate treatment based on race. And as far as I can figure this isn’t true. If anything, the numbers support the opposite conclusion.
As a non-math person I would welcome it if Hebegebe or any other math person explained to me what I got wrong.