A scientist can have an inclination towards--for example--racist ideas. You can't just call this a kind of being wrong, because depending on the truth of what they're studying, this can make them right more often or less often.

So racist scientists are possible, and racist scientific practice is possible. I think 'racist' is an appropriate label for the conclusions drawn with that practice, correct or incorrect.

Though, I think being racist is a property of a whole group of conclusions drawn by scientists with a particular bias. It's not an inherent property of any of the conclusions; another researcher with completely different biases wouldn't be racist for independently rediscovering one of them.

It's a useful descriptor because a body of conclusions drawn by racist scientists, right or wrong, is going to be different in important ways from one drawn by non-racist scientists. It doesn't reduce to "larger fraction correct" or "larger fraction incorrect" because it depends on if they're working on a problem where racists are more or less likely to be correct.