I think that depends on how severe they are. If that's 12 thousand reports of covered up beatings, that's extremely alarming. If it's 12 thousand reports of "showed up five minutes late to work", that's less alarming.
They're complaints by the public about police treatment of them, so have nothing to do with "showing up late to work".
They are also only records where the complaint has been substantiated by the [tilted towards police] Civilian Complaint Review Board, so only the most egregious and well-documented complaints are present. If you got the crap beat out of you, but no one else saw it and the officer says you fell down the stairs 8 or 10 times, that's not a substantiated complaint. If there's video footage of your beating but the NYPD declines to turn the footage over to the CCRB, then that's not a substantiated complaint.
The database includes records where the complaint is unsubstantiated or even exonerated, as long as the subject of the complaint has other complaints that were substantiated.
I think my point stands that some allegations are more serious than others and the aggregate seriousness of this database is a function of how serious the allegations in the database are. This seems almost tautologically obvious.
How would you measure the aggregate seriousness of a database, if not with some function of the individual entries of the database? Give me a fucking break, you know damn well I'm right.
