On the other hand, I must concede that I previously failed to look closely enough on first inspection at your third link. In my opinion, it does a great deal more to substantiate your statement than the IAT stuff does. I'd like to see similar studies with much higher n, but it's hard to argue with the analysis.