I wouldn't generalize this way. P( Y | X ) != P( Y ). X: Hired by google, good grades, good school. In fact, it doesn't test the difference between institutio

I wouldn't generalize this way. P( Y | X ) != P( Y ). X: Hired by google, good grades, good school. In fact, it doesn't test the difference between institutionalized tertiary education vs a lack of it, only differing academic performance among those who are already relatively successful in it.