What To Believe: The Polls or Horse Race News?

If you believe the polling numbers coming out of the battle for the White House and Senate it looks pretty unlikely Republicans will hold any branch of government come January.  Poll after poll spit out from non-partisan and partisan pollsters has shown Republican Senators in pink states trailing and the President running well behind nationally while also lagging (less so) in battleground states.  Indeed, the President trails by 8.5 points in the RCP average of polls.

The argument behind why the President runs better in swing states is well documented.  The thinking goes states like PA, MI, WI, FL, NC, etc. are redder than the nation according to past election results (PVI or Partisan Voting Index).  This is due to a combination of demographics and money the argument goes.  Trump’s approval in these states also dramatically outruns his national numbers explaining the tighter contest.

Generally, the larger margins the President trails by get the most coverage.  The media feeds off it.  Hence, when CNN showed the President down by 14 points or Fox showing Trump down by nine they jumped on his tweets about “fake polls.”  Ironically, this could be repeating the same mistake the media made in 2016 – believing the polls could not miss (in mass) and forgetting elections are made up of the people who turn out – not who respond to polls.

By this, I mean the polling results often conflict with what we hear on the ground in the same way they did in 2016.  Four years ago, there was unrest with both Trump and Clinton but a deep unrest in rural areas with the direction of the Democratic Party.  You still hear the same reports, scattered as they are, but they get lost in the mutterings of Trump losing support in the suburbs.

It should be noted after 2016, the media actually engaged in some self-reflecting on how they could have been so wrong about the election and resolved to talk with and understand why voters would put Trump in office.  They did this for a few months and then returned to their past tendency of covering elections like horse-races – focusing on every poll, every sound byte, every sensationalist story – and forgetting yet again voters across the country decide elections.  So, Trump and Republicans were soundly crushed in the suburbs but performed even stronger in other areas.

The horse – race coverage of 2020, as opposed to the polling coverage (well, basically, Trump is toast if the polls are right), has been exactly like 2016.  Focus on every sensationalist story, every campaign mistake, every word uttered and forget, yet again, outside the Beltway, opinions might not really have shifted.

Take for example the incompatibility of polling showing the President down by 13 points nationally (and a rally indicating the President sold 1 million tickets in Tennessee).  It is true most might not show and the area is pretty heavily Trump.  But, it shows a depth of enthusiasm for Trump the polls seem to be missing.  Meanwhile, the horse race is more about Belt-way talking heads trying to explain voter preference in polls they probably barely understand.

None of this is to say the polls are sure to be wrong or neither is the media’s coverage.  But, polling, despite FiveThirtyEight’s protestations (which has a vested interest in keeping the pubic trusting polls), has not had a strong track record of late and herding is not a new phenomenon in their circles.  It is quite plausible the media polling industry, and media, even if the President loses, will have a lot of explaining to do if the election is not a landslide and/or even if Republicans manage to retain the Senate (which is better odds right now than Trump being reelected).

Just goes to show, the industry has not learned much from 2016 and probably won’t from 2020 either.




Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s