General Discussion
In reply to the discussion: Why I don't like the term "white privilege" [View all]spooky3
(34,444 posts)Statistics 101.
Re: (a) There is a major tennis tournament played annually in Palm Springs. For years, Serena Williams has refused to play in it, because she and members of her family were treated so poorly, and in a racist manner, by some of the fans attending the tournament. Apparently some of the fans were not so swayed by her present social class and wealth that they were able to put aside their racism.
Re: (b) Class and race are somewhat correlated in the US, because of many complex factors. However, this correlation is far from 1.00, a perfect correlation. Therefore, it is fairly easy to separate the effects of actual class (or wealth, or income) and race (or of perceived class etc., and race), depending on how the researcher wants to define these terms, on whatever outcome variable interests you (let's say it is store clerk's speed of asking "may I help you?", but you can pick whatever you like). You are essentially arguing that the effects of class (etc.) would have a stronger weight in a regression analysis than would race (with other relevant variables controlled). You may even be arguing that race would have zero effect once class was entered as a predictor. Are you able to produce those studies?
A predictor does not have to have a 1.00 relationship with an outcome like the store clerk's speed in order for us to say that it is significantly related. Often there are multiple predictors that are significant. That means that multiple factors may affect an outcome.
And, only where there is a 1.00 relationship with the outcome would one expect that all whites were treated better than all blacks, for example. Social scientists NEVER observe 1.00 correlations in real life. I am not a race researcher, but I am aware that there are many studies that show race to have a significant (but not perfect) effect on various outcomes even when other variables are controlled or identical.
White privilege exists if being white is *significantly* related to whatever outcome we're discussing, once the effects of other important variables have been controlled ("holding everything else constant", "all other factors being the same", etc.). In a large sample, that coefficient does not have to be very high to be statistically significant, e.g., the regression coefficient for your SAT score relative to your college grade point average may be only .3, but with thousands of people in the sample, we can be very confident that that relationship in the population is greater than zero. But that is not the same as saying that SAT is the only important predictor of GPA.
So it is statistically ridiculous for people to refuse to acknowledge white privilege just because some, even many, white people are less well situated than some blacks.