If having such ancestors completely explains why Kenyans have that trait, then that would count as “the correlation between race and bad things is eliminated once you take the other factor into account”.
What other factor? If you mean “having such ancestors” then you’ve just tabooed the word “race” but I don’t see how this counts as “eliminating the correlation”.
If people of some race are more likely to have a bad trait, but that increased likelihood can be completely explained by the fact that people of that race are more likely to have some other factor, then there is no correlation between race and that trait once you condition on the other factor. That’s what “completely explained” means.
If that increased likelihood cannot be completely explained by the fact that people of that race are more likely to have that other factor, then there remains some correlation between race and that trait even after conditioning on that other factor.
What other factor? If you mean “having such ancestors” then you’ve just tabooed the word “race” but I don’t see how this counts as “eliminating the correlation”.
If people of some race are more likely to have a bad trait, but that increased likelihood can be completely explained by the fact that people of that race are more likely to have some other factor, then there is no correlation between race and that trait once you condition on the other factor. That’s what “completely explained” means.
If that increased likelihood cannot be completely explained by the fact that people of that race are more likely to have that other factor, then there remains some correlation between race and that trait even after conditioning on that other factor.
The problem is that in the grandparent the other factor is just a reformulation of “have the same race” in different words.