Thread by @ewzucker

ELECTION DAY THREAD w fresh results on Trump’s campaign for the midterms:

The NYT says that it was about stoking fear. But do voters agree? & what are the implications?

(1/21)

We (@ohahl, @minjaekim22, & @ewzucker) conducted a survey on Amazon Turk (N=630) over last week to shed light. Focus was on Trump’s 10/22 statement that the caravan “includes middle easterners.” Only American MTurkers outside large blue states (CA, MA, NY, NJ) cd take survey.

Idea was to get sufficient Trump supporters & indeed got 279 (40%) who were at least “somewhat” approving of Trump 72=11.4% expressed “strong approval.” By contrast, 247 (39.2%) strongly disapproved of Trump, & 343 (54.4%) at least “somewhat” disapprove.

So NOT technically a representative sample, but in the distribution of Trump approval & other characteristics (see graphic), sample looks broadly representative. A couple of notes before we get to the question of “fear.”

First, Trump-approvers were less aware of this “middle eastern” statement than Trump-disapprovers were. 94% of the Trump-approvers picked out the statement from two other statements (referencing “Colombians” and “Venezuelans”) as compared to 78.1% of Trump-disapprovers (t=3.32).

Second, Trump-approvers were more likely to say that the statement was true (mean of 2.77 on 7-point scale where 1 is “Factually True” & 7 is “Factually Untrue”) than were Trump-disapprovers (6.09; t=29.1). This is unsurprising of course. Note however the asymmetry:

Trump-disapprovers are a full point closer to complete rejection of statement’s truth than Trump voters are to endorsing its truth of the statement. This is in line with our post-2016 election survey (see appendix of journals.sagepub.com/doi/abs/10.117…)

& as with that survey, we found in this one that Trump-approvers are much more likely to agree that the statement should not be taken literally (mean of 3.8 on 7-point of scale of ‘nonliteralness’ vs. 2.0 for Trump-disapprovers; t=12.57), following Salena Zito’s famous quip.

We also asked “In assessing President Trump's statement that "Criminals and Middle Easterners are mixed in,"what is a more important consideration for you:that the statement should be based on objective evidence OR that it should send the right message about American priorities?”

Findings echo literalness findings: Trump-disapprovers say it shd be based on objective (mean of 2.2 on 7-point scale from “It should be based on objective evidence” to “It should send the right message on American priorities” vs. mean of 4.5 for Trump-approvers (t=15.54).

But here’s the thing. While Trump-approvers are less focused on what Trump literally says & they ascribe less importance to whether it’s objective than Trump-disapprovers, there's broad agreement w the NYT (& the “fake news” media generally) on the content of Trump’s message.

Consider answers to 2 Qs. They were asked if caravan statement was meant to a) “scare the public." They were also asked if statement was meant to send "message that illegal immigrants are a problem.” On avg, both Trump-approvers & Trump-disapprovers tended to *agree* with both.

To be sure, Trump-disapprovers expressed greater agreement with “scare people” (Mean of 6.37 vs. 4.49 for Trump-disapprovers on 7-point agreement scale; t=15.1).
Meanwhile, Trump-disapprovers expressed greater agreement as to “illegal immigration” (Mean of 5.83 vs. 5.29; t=4.03).

Broadly though, the majority of both 'sides' agree as to content of Trump’s message even if there's disagreement as to whether to take his words literally or to care about objective foundation.
Ah, but here’s the last thing to note: there is *sharp* disagreement as to…

… how much they *like* Trump’s message. Overall & unsurprisingly, Trump-approvers are more likely to agree they're “happy” Trump made the statement (mean=5.16 on 7-point scale vs. 1.71; t=33.53). More interesting is how recognition of Trump’s message correlates w happiness...

For Trump-disapprovers, the more they see Trump as “scaring” people, the more unhappy they are (r=-.48). By contrast, the association is slightly *positive* for Trump-approvers (r=.06). That is, Trump-disapprovers are very upset that Trump is trying to…

… scare the public but Trump-approvers aren’t bothered by it. Conversely, while Trump-disapprovers are happier the more they see Trump’s message as identifying illegal immigration as a problem (r=.25), this association is mildly negative for Trump-disapprovers (r=-.05).

Upshot: Despite their different orientations towards (the importance) of facts/objectivity, it appears that Trump-disapprovers and approvers (at least in our sample) see roughly the same Trump (same as “fake news”/MSM media), but it resonates very differently with each 'side.'

As to *why* the two sides have such diff orientation to facts & message, we've developed what we think is useful framework w evidence (both from 2016 election aftermath & from a simulated context)

(Our application to Kavanagh seemed to resonate. See

https://twitter.com/ewzucker/status/1046219643407470592

)

But other research also sheds important light & more research is needed.
In that respect, please take the results reported here as *very preliminary.* They are very fresh and this has not yet gone through peer review-- shared only bc timely.

Please send feedback to...

... the research team (who contribute *equally* to this thread as well as papers!): @ohahl, @minjaekim22, & @ewzucker. THANKS!

/FIN

P.S. Don't forget to vote! And let’s reach out to the other ‘side’ from us after the election!