Not having read Doctorow's post, I might have picked up the wrong detail, but it sounds like they are attempting to summarise multiple dimensions into one. Because sure, if one tskes a highly complex data set, runs factor analysis, and constrains it in a particular way, you're going to get a factor that highly correlates to several measures. In this case that bigotry against 'other' is 'traditional'.

I'm not quite managing to make that coherent, but my inner psych/statistician thinks that Doctorow is torturing the data without adequate justification.