The more slave-intensive a southern county was 150 years ago, the more conservative and Republican its contemporary white residents. The authors tested their findings against every plausible control factor—for instance, whether the results could be explained simply by population density—but the correlation held. Higher levels of slave ownership in 1860 made white Southerners more opposed to affirmative action, score higher on the anti-black-affect scale, and more hostile to Democrats.
From Jonathan Chait's article in
Edit: Oops on the magazine name; thanks to Anon. for correcting me.