I started writing a take on the national and state polls as a way to calm myself down and sent it around to friends and family. People seem to like it, so I’ve moved it to Substack and given it a name - Pollapalooza. Feel free to pass it along, and people can sign up via the button at the bottom of the email.
OVERVIEW
Two months out from election day, the numbers look very good for Biden, which is reflected in the analysis below, but I also think we are in for a hellish couple of months.
The Republican National Convention wrapped up late last week and I’ve been waiting for polls that were in the field after the convention. And yesterday was pollapalooza (the new name of this email) with 10+ new polls released including Fox, Morning Consult, CNN/SSRS, Quinnipiac, Ipsos, and YouGuv. The headline is Biden is leading nationally in all of them by an average of +7.2%. That is really good. It’s only .1% off of where he was before the conventions. If an incumbent is down, the convention is one of the only events that can change the narrative to improve their position. Trump failed.
STATES
It’s good to be up nationally, but we are all painfully aware the Electoral College is how the US elects presidents. Biden is holding a significant EC lead to match his national lead. I’m going to highlight swing states where we got poll results in the last two days, highlighting the Morning Consult and Fox News polls. First, from Morning Consult, a comparison of their pre-convention and post-convention poll results in the swing states. https://morningconsult.com/2020/09/01/battleground-presidential-polling-post-conventions/
The only dramatic swing was in Arizona, where Biden improved from 45% pre-convention to 52% post-convention. Fox also had a poll out yesterday showing Biden with a 9pt lead in AZ post convention. It is worth noting that in the FiveThirtyEight polling averages, Trump has not led in Arizona since early March. There is also a key senate race in AZ between Republican Martha McSally and former astronaut and Gabby Gifford’s husband Mark Kelly. The Fox poll has Kelly beating McSally by a whopping 17pts. We may see an up ballot bump to Biden in AZ because of the McSally/Kelly race.
https://static.foxnews.com/foxnews.com/content/uploads/2020/09/Arizona.pdf
Also, notice that Morning Consult has Trump up by only 1pt in Texas post-convention, AND they only had him up by 1pt pre-convention. FiveThirtyEight’s polling average has Trump leading Biden by 1.7% as of today, and it has been that tight since the middle of March. Is Texas in play? Under normal circumstances it would be hard to argue it isn’t. However, Republicans will suppress the vote and the pandemic is a variable we’ve never had to consider in turn out models before. I’m going to act like it is and work on voter turnout. Remember, Beto only lost by 2.6% to Cruz, over performing projections.
I can hear y’all rolling your eyes at the mention of Fox News, but I’m telling you their polling shop is top notch. I know that sounds crazy, but historically they have not pulled any punches with their polling. It is one of the polls I trust the most.
In addition to Arizona, Fox polled North Carolina and Wisconsin post-convention. They are three states Clinton lost that Biden could claim.
Arizona: Biden up 9pts
Wisconsin: Biden up 8pts
North Carolina: Biden up 4pts
Notably, in both Arizona and Wisconsin voters favor Biden on policing and criminal justice by 5pts. This poll was in the field after police shot Jacob Blake and the data point gives me some hope for humanity. Biden and Trump are tied on policing and criminal justice in North Carolina.
FUN WITH MAPS
Neither of the two maps below reflect where I think we’ll be on election day, but they are good for visualizing how hard the race is for Trump right now.
In the first one, I started with the 2016 map and moved the states that were red in that race to blue based on the Morning Consult poll. In that scenario, Biden wins the EC 284 to 254. Link to interactive map.
In the second one, I matched today’s state poll averages at FiveThirtyEight and categorize states:
- Poll averages below < 2% as toss ups, green
- Poll averages between 2% and 5% as lean, light blue or pink
- Poll averages between 5% and 10% as likely, blue or red
- Poll and > 10% as safe, dark blue or dark red
The map is starkly awful for Trump. Biden wins with 319 EC votes to Trumps 115 EC votes, with 104 EC votes in toss-up states. That means, Trump has to run the table on toss-up states AND win 51 EC votes from states that lean Biden as of today - two months out from election day. That is an enormous hill to climb. The Republicans are going to cheat in every imaginal way, so it’s not impossible. But it would be without gaming the system. Link to interactive map here.
Tools
FiveThirtyEight’s election poll averages. The best source for daily poll updates, broken out by national and state.
https://projects.fivethirtyeight.com/polls/president-general/national/
The Financial Times presidential election tool. I really like the layout and they pull from Real Clear Politics poll averages, which lean more republican, helping me to avoid confirmation bias.
https://ig.ft.com/us-election-2020/
Real Clear Politics includes pollsters that FiveThirty Eight considers too republican partisan and excludes from its model, but still worth looking at. https://www.realclearpolitics.com/epolls/2020/president/us/general_election_trump_vs_biden-6247.html
270 to Win has an easy interface to build maps and it is where I plug in information I pull from the other sites.
https://www.270towin.com/