Yes, despite that you are grayed at this point, that example is not logically the same as the phenomenon under discussion.

If you talk to a lot of people with kidney failure you might conclude kidney failure is more likely than it is. This makes no sense as you are specifically looking for people with kidney failure, but humans aren't great at perspective here.

If you are looking for murder victims and a separate attribute--where these murders take place--is clearly correlated with murder victims, it is correct to notice that correlation.

You can say that a neighborhood has more shootings/murders than other neighborhoods. What I'm talking about is the phenomenon of thinking of that neighborhood just in terms of those incidents, because that's all you ever hear about them. You aren't there for the 99.9% of the time when things are mundane.