1/n To what extent did anti-PC sentiment influence voting in the recent midterm elections? Absent the relevant survey data, there's no way to know. All I can say is that it did *appear* to influence (if only modestly) voting for GOP Congressional candidates in the 2016 election.
2/n To show this, we'll begin with a basic logit model, which regresses votes for GOP (Senate and/or House) candidates on measures of anti-PC sentiment, economic pessimism, and the standard series of demographic/political controls. 
3/n As shown, net of all other predictors, anti-PC sentiment significantly predicts voting for a GOP Congressional candidate (OR=1.47; p < 0.001).
4/n More specifically, moving from the bottom (pro-PC) to the top of the anti-PC scale coincides with a roughly 25 % point increase (54.5%->79.3%) in the probability of voting GOP.
5/n Interestingly, the same shift sees similar effects even among those who identify as 'very liberal'/'strong Democrat' (Predicted Probability= 6.8% [pro-PC]-->18.9% [anti-PC]). In fact, it appears (perhaps naturally) that the effects are most influential for those who place...
6/n ...at the more moderate points along the political orientation scale (in other words, strong conservatives/Republicans vote GOP regardless). Though voting for Trump may have been a bridge too far for them, anti-PC Clinton voters were also somewhat more likely to vote GOP
7/n To test the robustness of this relationship, let's now add a few more election-related variables to the model; namely, anti-immigration attitudes, 'modern sexism', and 'racial resentment'.
8/n As shown, the inclusion of these variables does indeed moderate the effects of anti-PC sentiment. Nevertheless, the latter remain significant (OR=1.28; p=0.013).
9/n Holding all other variables at their medians, moving from the min to max of the anti-PC scale coincides with a 15.4 % point increase (61.6%-->77%) in the odds of voting GOP
10/n Thus anti-PC sentiment does appear to have been at least modestly relevant to GOP votes in 2016. Whether it mattered in some of the races on Tuesday, though, is an open question.
11/n For the Dem Party in 2020, these results suggest that fielding outspokenly 'activist' prog. candidates carries some degree of risk (as many pundits have warned). Although they may mobilize some important constituencies, they are likely to dampen enthusiasm among others.
12/n For instance, among Clinton Voters, those who think PC has gone too far were significantly 'cooler' ('warmer') towards Clinton (Trump) and were sig. less likely to characterize their vote preference as 'strong'.
13/n Furthermore, recent research suggests that candidates who use 'filtered' or PC-speech are perceived as less authentic--irrespective of candidate preference. onlinelibrary.wiley.com/doi/abs/10.111…
14/n And, incidentally, in the present data, the gap between how Clinton voters assessed Clinton vs. Trump's honesty was significantly, if slightly, smaller among the anti-PC
15/n FYI, this topic will likely serve as the basis for some of my dissertation. I plan to conduct an experiment that attempts to replicate Conway et al. (2017) jspp.psychopen.eu/index.php/jspp…. So stay tuned--plenty more on this to come :)
The End.
The End.
