I got this impression as well, especially given >> Participants’ reported political affiliations were 40% Democrat, 20% Republican, 36% independent, and 4% oth

I got this impression as well, especially given >> Participants’ reported political affiliations were 40% Democrat, 20% Republican, 36% independent, and 4% other I think that distribution is quite a bit different than a random sampling of US adults would get you. For a study examining political preference, having such a bias is pretty harmful. Perhaps their effect would have been more pronounced with more Republicans, or maybe it would have been less pronounced. Maybe the unusually high percentage of independents dampened the precondition of extremism, or maybe it actually exacerbated the precondition.