The First Pancake, Part One

Hi all,

Over the last few months, I did two big public-facing pre-election projects. This post is about the first one: A series of seven House polls we did with our friends at Inside Elections.

The basic idea was simple — poll a variety of competitive House districts, using short scripts and phone based methods. This was great for a few reasons:

• We got to poll places and races nobody else was looking into 

• Inside Elections is really easy to work with  

• We had a quick feedback loop. If someone objected to something about our approach, we could hear them out and integrate helpful thoughts into the next survey.  

Here are the basic results of each survey (some of this is off memory, apologies if I got something wrong!) 

MI-07 (Lansing): R+7. 

• This was at the height of Trump hype, right after Biden blew it in the debate

• Republican enthusiasm (and response rates) were through the roof. Had to adjust for that.   

• I think our numbers reflected reality at the time – which was a big GOP advantage – but things have calmed down since. 

OH-09 (northern Ohio/Toledo): D+10. 

• This was just a little later – the height of Harris-mania. Crazy Democratic response rates 

• Candidates matter: Clearly Marcy Kaptur’s blue dog Democrat thing is helping 

• I would be shocked the results were this good for Democrats – it was a low point for GOP enthusiasm too 

OR-05 (NW OR, not much of Portland): D+2. 

• A lot of undecideds were Democrats. If I had to guess, Democrats do better than this poll suggests in the end. 

• It was fun to get the lay of the land in a state that receives almost no national attention!

MT-02 (Western Montana): R+4. 

• Not a surprising margin in the House. Horrible numbers for Tester (tied) in the Senate race. If he can’t win MT-01 by a solid margin, he can’t win statewide. 

• If I had to pick one poll where I think our numbers have the best chance of holding up, it’s this one. I’m jinxing it, but that’s how I feel. 

NC-01 (Northeast NC): D+8 

• We did this one during the hurricane. It was a nightmare. Response rates fell off a cliff in the middle of fielding and I was super nervous our sample was wacky/bad 

• Then I got less nervous when a Dem firm (expect them to be about ~3 points better than reality for their side) came out with a D+11 poll. So right in line with us, after adjusting for that.  

• Things might have shifted since then. But maybe this one turned out okay? TBD. 

AZ-02 (Northeast AZ, but kind of winds around the state weirdly): TIE. 

• This won’t be a tie. It was Trump+9 on the presidential line, and there were a ton of undecided Republicans in the House race. 

• BUT the Democrat – Jonathan Nez, former president of the Navajo Nation – might drive up Native American turnout. If he does, I think this result anticipated that strength 

• It was probably late enough in the cycle to warrant pushing undecided voters to just pick someone. Maybe we should have. But you can’t change the past! 

PA-08 (Scranton-ish): D+7. 

• By this time, we had done a lot to integrate feedback – on weighting, collection practices, etc. I feel good about all our polls, but this was our best effort. 

• And honestly, it put us out on a limb! It’s a really good result for Democrats. We got Trump+3 in the presidential race and a Dem winning the House seat by seven. That’s a big gap! 

• TBD if all that ticket splitting holds up. We will see! 

Most of these polls are too far off from the election to be judged as predictions (except PA-08). But it is reassuring – from a “did we do a good job on this” perspective – that there’s a mix of good results for Dems, good GOP results and ambiguous numbers. 

Also – almost forgot to explain why the title is “The First Pancake!” 

I got this phrase from Kristen. Basically, when you make a batch of pancakes, the first one is always a little lopsided, overcooked, weird – just off in some way. But that’s the nature of making pancakes. You accept that it’s the first pancake, learn from it and improve on the next dozen. 

I’m terrible at this mindset. I never forgive or forget any of my mistakes. And I worked hard on these polls! I want them all to look like little pearls of prescience by late Tuesday night. 

But I can’t control that. This is my first election cycle out of journalism and in polling. So I’m trying to integrate First Pancake Thought into the brainspace. It’s more honest than pretending to be all-knowing. And, if you’re putting something out there publicly, this mindset is probably the only way to survive and grow. 

Thanks for reading! 

David