The numbers can be skewed for so many reasons that it would hard to list them all. Thanks for mentioning this, however; I considered doing so in the article but then just didn't.

But as to these three markers for racism, it is the percentages of each group, 1) non-Hispanic whites and 2) other racial and ethnic groups that is considered. So, for example, if 75 percent of the first group graduates from college but only 25 percent of the second group graduates, that indicates a real disparity in education. It isn't the numbers that is being compared, it is the percentage of people in each group.

As to the averages, you are right that resulting numbers can sometimes be skewed. When you're working with numbers from 1 to 50, as I was here, there are no huge or miniscule numbers that would skew an average. But I do try to stick to mediums where the numbers are like your example--since mediums ignore both top and bottom figures that would skew results.

Thanks for your comment, overall. Do keep an eye on me and point out where I've missed something. I appreciate the criticism.