[ketralnis]

> There are around 700,000 police officers in the US. If one of them is caught doing something bad every day, that's 0.00014%.

I take your point, but not exactly. First, an officer that's planting evidence or attacking unarmed handcuffed civilians has been known to be a bad officer by say the other 50 officers in their department. Second, they're not caught on their first day. Third, the high profile ones are high profile because they're left on the force and continue being bad even after being caught.

So the 0.00014% of people being caught every day is _additive_ over time, this isn't a random sampling. It doesn't mean that 0.00014% of officers are bad, it means that we increase the pool of bad officers by that much every day.

[antonvs]

The number was just intended to indicate what a small fraction any individual instance is. Even being additive, and even with undiscovered bad cases, these cases are still not representative samples of US police in general. In other words, as an earlier commenter said, this is partly confirmation bias.

That's not to say these incidents aren't indicative of systemic problems, but that's not the same as saying saying all or most police are like this. What "confirmation bias" means here is that people are looking at the exceptional cases to confirm their opinion of police as a whole, which is objectively a mistake.