And that evidence is? The IAT researchers' bibliography goes back to the mid-'80s. Have they found any correlation between changing unconscious associations and changing socioeconomic disparities over the past 20+ years?
I must assume you are referring to the bibliography of Blindspot: Hidden Biases of Good People, although perhaps not, as the publishing dates of sources likely relevant to your query go back to at least the 1920s -- ex. Bogardus (1928, 1925). If you haven't accessed this bibliography I heartily recommend that you do, as cursory examination of the sources included therein reveals a multitude that would likely address both your first and second challenges.