Up to a point. While you focus on the example the "literal retard" that's not where the problem comes in. The problem here is that people use the evidence more strongly than they should. Essentially this is the Screening off evidence problem. If I know say someone's standardized test scores and GPA that will be overwhelmingly more useful for predicting how intelligent they are than any weak prior based on race, gender or socioeconomic class. But humans often don't act that way.