That just becomes a question of how narrowly you group peopple. Murderers commit 100% of the murders.
That just becomes a question of how narrowly you group peopple. Murderers commit 100% of the murders.
right, but if toggling one attribute (race) moves the output more than others then you need to consider it.
I'm guessing that this could be applied to things like "HS graduate B-average" would have a similar effect. The main thing is that they use "gun crime" as a political tool to attack "white people, with long guns, who aren't leftists", when the truth is that it is black/brown golumn in cities (leftist voting) that are where the stat comes from.
right, but if toggling one attribute (race) moves the output more than others then you need to consider it.
I'm guessing that this could be applied to things like "HS graduate B-average" would have a similar effect. The main thing is that they use "gun crime" as a political tool to attack "white people, with long guns, who aren't leftists", when the truth is that it is black/brown golumn in cities (leftist voting) that are where the stat comes from.
But that's not a correlation that imparts predictive information. A correlation between two or more variables gives predictive power.
But that's not a correlation that imparts predictive information. A correlation between two or more variables gives predictive power.
(post is archived)