Comparing National Polls in the Presidential Election

People who don’t like Donald Trump often point to his refusal to accept the outcome of the 2020 Election as valid, but there is a reason for Trump’s concern.  If you look at both the 2016 and 2020 election results, Trump’s share of the vote improved in thirty-three out of fifty states from 2016 to 2020, as did his share of the popular vote nationally.   It seems incredible to believe you can improve in so many places, yet lose.   The difference between 2016 and 2020 can be seen in context, however, by seeing Biden’s share of the vote in 2020 versus Clinton’s share in 2016.

 

Biden’s share of the popular vote in 2020 was higher than Clinton’s chare of the popular vote in 2016 in every state and District … every one.  People can (and have) argued whether such an outcome reflects an impressively efficient campaign, or implies an outrageous level of corruption in the process, but the point here is that the Democrat candidate’s numbers  may well drive the outcome in 2024, so the best indicator of the present election is to check Harris’ support numbers.

 

I will not waste time here looking at Biden’s numbers – it was clear that he was irrelevant to the election before the summer started.  His last chance ended with the June debate performance against Trump.  That does not mean that the Democrats have run a clean, efficient campaign this year, however.  The timing of Biden’s recusal from the election, Harris’ inability to properly vet her VP nominee, and the clumsy media confusion during June and the first half of July demonstrate that while the DNC was aware of Biden’s weak position and was considering replacing him, concerns about Harris’ ability to win prevented the party from preparing a comprehensive campaign for Harris.

 

That lack of preparation is reflected in the present state polling.   Harris’ poll numbers at this time are lower than Biden’s 2020 election results in forty six states and three districts.  Her national numbers are also significantly lower than both Biden’s final election results and his polling numbers at this point in the 2020 race.  This loss of support has not been reported in the media, despite its implied effect on the 2024 condition.   Instead, the media has been selling the race as ether neck-and-neck, or saying Harris has a narrow lead after a prolonged media blitz selling all of Harris’ presumed advantages and none of her mistakes or issues.  2020’s results show that such tactics can potentially help a media-preferred candidate, while 2016’s results show that such tactics can damage a campaign, if valid concerns by voters are ignored by a major candidate.

 

So that brings us back to the current polling, and what the numbers say.  For simplicity, let’s start with the National polling, and note the same polls’ accuracy in the 2016 and 2020 elections.

 

There are three major polling aggregators out there:  RealClearPolitics (RCP), 270towin (270), and FiveThirtyEight.com (538).   They published results from the following polls:

RCP:

CNBC

Rasmussen Reports

NPR/PBS/Marist

SurveyUSA

Morning Consult

CBS News

I&I/TIPP

Marquette

Daily Kos/Civiqs

Harvard-Harris

Reuters/Ipsos

Wall Street Journal

NY Times/Siena

 

270:

 

RMG Research

CNBC

Marquette Law School

Economist/YouGov

NPR/PBS/Marist

SurveyUSA

Morning Consult

UMass Amherst

TIPP Insights

CBS/YouGov

Clarity Campaign Labs

American Pulse

Civiqs

Economist/YouGov

Redfield & Wilton

 

538:

 

Quantus Polls and News

RMG Research

Ipsos

YouGov

Activote

SurveyUSA

Morning Consult

Marist College

CNBC

(italics notes polls cited multiple times)

 

So that gives us the following polls for our use:

 

Activote

American Pulse

CBS News

Clarity Campaign Labs

CNBC

Daily Kos/Civiqs

Economist/YouGov

Harvard-Harris

I&I/TIPP

Marist College

Marquette

Morning Consult

NPR/PBS/Marist

NY Times/Siena

Quantus Polls and News

Rasmussen Reports

Redfield & Wilton

Reuters/Ipsos

RMG Research

SurveyUSA

TIPP Insights

UMass Amherst

Wall Street Journal

YouGov

 

But we still have more weeding to do.   To know how reliable a poll is, we need to see how it did in 2016 and 2020, and not all of these polls published in those election years.   Here’s how they did, comparing their Final polls to actual election results:

 

Activote: did not publish in 2016 or 2020

American Pulse: did not publish in 2016 or 2020

CBS News: 2016  and 2020 polls

Clarity Campaign Labs: did not publish in 2016 or 2020

CNBC: 2016 and 2020 polls

Daily Kos/Civiqs: 2016 and  2020 polls

Economist/YouGov: 2016 and 2020 polls

Harvard-Harris: 2020 polls

I&I/TIPP: 2016 and 2020 polls

Marist College: 2016 and 2020 polls

Marquette: did not publish in 2016 or 2020

Morning Consult: 2016 and 2020 polls

NPR/PBS/Marist: 2016 and 2020 polls

NY Times/Siena: 2020 polls

Quantus Polls and News: did not publish in 2016 or 2020

Rasmussen Reports: 2016 and 2020 polls

Redfield & Wilton: 2020 polls

Reuters/Ipsos: 2016 and 2020 polls

RMG Research: 2020 polls

SurveyUSA: 2020 polls

TIPP Insights:  2020 polls

UMass Amherst: 2020 polls

Wall Street Journal: 2020 polls

YouGov: 2016 41-46 Clinton, 2020 polls

 

 

 

So with that understood, here are the polls which published in either 2016/2020, with average results set against their most recent numbers (August 11) and subsequent adjusted numbers (Trump vs poll = T+/-, Democrats vs poll = H+/-):

 

CBS News: T+6.2, H+1.3  adjusted 51.2 to 47.3 Trump

CNBC: T+8.5, H+2.8    adjusted poll 53.5 to 45.8 Trump

Daily Kos/Civiqs: T+3.0, H+3.3  adjusted 47.0 to 51.3 Harris

Economist/YouGov: T+4.5, H+0.8  adjusted 47.5 to 45.8 Trump

Harvard-Harris: T+1.9, H-3.7  adjusted 50.9 to 48.7 Trump

I&I/TIPP: T+1.0, H+3.3  adjusted 46.0 to 49.3 Harris

Marist College: T+3.5, H+0.8  adjusted 49.5 to 49.8 Harris

Morning Consult: T+3.5, H+1.3  adjusted poll 47.5 to 49.3 Harris

NPR/PBS/Marist: T+3.5, H+0.8  adjusted poll 51.5 to 51.8 Harris

NY Times/Siena: T+5.9, H+1.3  adjusted 51.9 to 46.3 Trump

Rasmussen Reports: T+0.5, H+4.3  adjusted poll 49.5 to 48.3 Trump

Redfield & Wilton: T+5.9, H-1.7  adjusted 49.9 to 43.3 Trump

Reuters/Ipsos: T+6.0, H+3.3  adjusted poll 48.0 to 46.3 Trump

RMG Research: T+2.9, H+0.3  adjusted 48.9 to 47.3 Trump

SurveyUSA: T+2.9, H-0.7  adjusted poll 47.9 to 47.3 Trump

TIPP Insights:  T+1.9, H+1.3  adjusted poll 46.9 to 47.3 Harris

UMass Amherst: T+2.9, H-1.7  adjusted 45.9 to 44.3 Trump

Wall Street Journal: T+4.9, H-1.7  adjusted 53.9 to 45.3 Trump

YouGov: T+4.5, H+0.3  adjusted 49.5 to 45.3 Trump

 

19 polls, aggregate average is 49.1% Trump, 47.4% Harris.  A modest lead for Trump and within the statistical margin of error, but note that when adjusted for historical performance the numbers are better for Trump than the polls currently indicate, and also that this is the first of three elections where Trump has held and maintained a lead in the aggregate national polling.

 

Winners Wizbang Caption Contest™ Week of August 9, 2024
Wizbang Weekend Caption Contest™ Week of August 9, 2024