This is really fascinating and fits with work I've done in the past questioning the value of self-rated prejudice measures to track trends in prejudice...1/?
Trends in the Q "do you consider yourself prejudiced against people from other races?" are basically flat over four decades. This is puzzling because other more concrete questions (e.g. Views of inter-marriage) show sharp decline. Resolution to the puzzle is changing standards.
People who come from parts of British society where blatant prejudice is now rare (young, graduates etc) seem to apply a much tougher standard when rating themselves than people who come from groups where blatant prejudice more common (older, less educated).
I was unsure what the mechanism was for this - the blog linked to provides a v plausible candidate - people's definitions of a phenomenon depend on the frequency of its occurrence
I blogged about the self-rated prejudice measure a while back here: blog.policy.manchester.ac.uk/featured/2015/…
