A Solution For Crowded Primaries – David Kanevsky (3D Strategic Research)

David Kanevsky:

Even the candidates who don't win end up with positive impressions among voters that lets them run for future offices.

Eric Wilson:

I'm Eric Wilson, managing partner of startup caucus, an investment fund and incubator for Republican campaign technology. Welcome to the business of politics show on this podcast, we bring you into conversation with the entrepreneurs who build best in class political businesses, the funders who provide the capital and the operatives who put it all together to win campaigns. Our guest today is David Kaneski of 3d strategic research. He's a data driven strategist with a long career at organizations like the national Republican senatorial committee and the Republican state leadership committee, as well as experience as a campaign manager on the ground. David and I recently collaborated on a poll in two Virginia congressional districts to understand some of the effects of rank choice voting in primary contests. In our conversation today, we discussed that poll, what we learned and what Dave sees as the future of polling David. We had a unique opportunity to do what is kind of a rare AB test in Virginia's seventh and Virginia 10th for this survey. What were some of the similarities and differences that made these two districts worth testing out?

David Kanevsky:

So most AB tests on campaigns really focus on kind of micro tests. What's the impact of voter contact, you know, oh, this turnout piece or digital with the control group to see, okay, here's the people who got it. Here's the people who didn't, but it makes it really difficult to measure big things. Like how do you measure candidate quality? How do you measure election systems? Because obviously, you know, there's not always a control group. You know, there's a lot of unique factors, but here what we did by having two districts, it's not a pure AB test, but it's probably as close as possible that when it comes to measuring rank choice voting, because one district had it, one district didn't, but the districts are really similar. One they're neighboring districts in Virginia, they're both in suburban and exurban Virginia, both are in the DC market.
So they kind of get covered by, you know, DC television stations. If candidates are advertising they're advertising on DC broadcast, both districts had multi candidate primaries to go up against incumbent Democrats who flipped districts in 2018 that had actually been Republican held districts before that. So again, you're seeing a lot of similarities about who the districts are. Their geography both had very crowded fields in the 10th district. We had 11 candidates running that was the district with rank choice voting. And in the seventh district, we had six candidates running. So again, a lot of similarities, there were a little, a few differences. A 10th had a firehouse primary that was run by the party. And that was the one with rank choice voting while the seventh had a state run primary with just the first choice. And so our goal with this kind of AB test wasn't to kind of get what the ballot was because we have that from election results. The reason why we did the survey was to look at what were the attitudes of the voters, one about the election, about the process they just went through and then look at how they viewed the candidates, look at how they viewed the process to see how do these voters think about rank choice voting and the candidates and see what impact rank choice voting had, where voters were more exposed to

Eric Wilson:

It. And just a little bit more context for our listeners. So in, in Virginia, the different congressional districts parties have the ability to determine their method of nomination. So they could do a convention, they could do a, a firehouse primary, or they could sort of hand it over to the state and, and do the, the typical thing that, that you're used to obviously conventions have been going on for a long time in Virginia but with the rise of COVID and, and restrictions on in person gatherings, the, the Republican party in Virginia started looking at rank choice voting in these sort of unassembled conventions or firehouse primaries as a way to preserve some of those mechanisms. And so it's been a really fascinating development scene on the ground. And, and of course we, we got a really cool opportunity to see what impact that has in these two neighboring congressional districts. Dave, one of the challenges we faced was finding voters in the 10th district because it was a party run process, which typically brings in a smaller pool of voters because you don't have as many locations, you don't have that mail-in ballot component. You don't have the early voting as widespread. And of course we don't have party registration in Virginia. So walk us through how you tackled that to get the sample size that you needed.

David Kanevsky:

So usually when you're trying to figure out who's a, a voter, whether it's a general election or a primary voter, you look at kind of what is past behavior cuz past behavior is indicative of future future behavior. The problem was here, there was really no kind of good precedent for what this election looked like. Because we don't always have party run firehouse primaries in Virginia, particularly with redistricted. You know, there were parts of the tent that were new. And so for example, if we had just looked at everybody who voted in the 2021 governors unassembled convention, we would've only had about 5,000 voters who cast a ballot in that in the 10th district. And what ended up happening was there were about 15,000 people who voted in the 10th firehouse primary this year. So if we had kind of looked at just that universe, would've been a little too small.
so what we looked at was who were people who had voted in past Republican primers, we started with that, but then we also looked at who were some new registrants who might be some low propensity voters and added them into the sample mix to make sure we kind of caught them because those people may be different. And that's really important, I think for firehouse primers because each candidate might bring their own unique friends and networks, even if they're not past Republican primary voters. So we haven't had the voter file come back yet in either of these districts, but wouldn't be surprised if, you know, there were a lot of first time Vietnamese voters who showed up for hung cow because that was kind of his network or Janine Lawson might have brought in people from her supervisor district who know her, who haven't voted in a Republican primary before.
So we were able to kind of use that to kind of capture these people. And we did something similar in the seventh. And what we do see there is in the seventh, again, we don't have the full voter file back, but we do know from the absentee data, we are seeing a lot of new primary voters coming in there. In fact there were twice as many people who had never voted in a Republican primary who voted absentee than people who voted in 4 0 4. It wasn't Dem crossovers. It's really just a lot of frequent general election Republicans who were starting to come in and that's really important to get a good sample. And it actually worked. So one of the things when I was at committees I always asked for not just, Hey, what was the top line data, but look at your unweighted data.
And here we actually see that when we look at the ballot unweighted versus what the actual election results, they were pretty close, they were all within, you know, two to four points. Well within the margin of error of what the actual election results were. So we did a little bit of waiting to get to the actual election, but had really no impact on the data, on RCV, things like that. But I think that's really important, not just for this survey, but if you're running for office, you know, yes, you should talk to, you know, high propensity voters, cuz they're very likely to vote, but you need to watch out for these low propensity voters cuz they really matter in a lot of these races, particularly close races.

Eric Wilson:

Yeah. And we ended up having a pretty large pool to survey from because you had roughly 15,000 people participate in that firehouse primary. And then over in the seventh it was around 30,000 give or take. And so we actually had some, some, you know, it was off by a factor of, of two, but not what you would typically see with like a convention where it would be really small. And, and Dave, I wanna underline something that you, you mentioned there, which is came up in my conversation with John Black, from the data trust in a recent podcast episode, which is that you've gotta be very, very careful if you're running a campaign and you're just targeting people based completely on vote history because to your point, the campaigns would've missed out on these people that might have come from a neighborhood network or a church network, if they were only doing their voter contact people with vote history.

David Kanevsky:

And actually if I can elaborate a little bit more on this, cuz I think it's actually really relevant. The seventh district used to be Eric Canner's district. And I actually did the polling for Dave Brad after he beat canner for the general election in 2016. And obviously that was a huge shock. No one saw that coming. I don't think Brad's team even really saw that coming as well. And people were like, what happened? Cuz can's pollster had him up like 30 some points, but what was the miss there? And it wasn't just there. It was in 2016, you know, Trump in the in the primaries turnout was way up with a lot of new Trump voters coming in the error we're seeing really in polling these days, isn't asking people who they're voting for. It's figuring out who's gonna vote to your point what John Black said.
And again, it's in with in Canter's case, it was, oh, we're gonna look at people who vote in the last primary. And when turnout went up 50%, they missed a lot of these new voters. And when we went back and looked at who were these new voters was they weren't Democrats, they were largely software Republicans and independents who really didn't like Eric canner and you know, came in and that's what changed the electorate. There another example where we saw big polling error and really people aren't talking about that as in Georgia, the public polls at the end had Kemp up about 20 and he won by 50 and yes they were directionally, right. But if the polls had Purdue up 10 and Kemp won by 10, everybody would be like, well, the polls are off. They were off 20 points here, even though they were direction off a 30 point error. It still really important. The reason was because turnout was way, way up again. We saw the same thing based on the absentee early voting a lot of new primary voters. So again, if you're not really capturing these low propensity voters making sure they're in the sample, like that's really, what's throwing off error and be interested to see, you know, the cycle from media polling, public polling, you know, how, how that shakes up. Because again, we may be onto another very high turnout midterm electric.

Eric Wilson:

So back to our poll in Virginia, one of our, our first finding was that rank choice voting produced a better positioned nominee in terms of net favorable image among voters talk about how big that gap was and, and what you think contributed to it.

David Kanevsky:

So in the seventh district where there wasn't rank choice voting yes. Vega who ended up winning, she had a good image Annette plus 51 on her favorability. So she had 64% favorable, just 13% unfavorable, but then we let's turn to the 10th district, hung Cal won that through rank choice voting, his net image was plus 78. He had an 86% favorable rating among people who voted in the Republican primary. They're just 8% unfavorable. So Cal is at plus 78. Vegas is at plus 51, that's a net 27 point difference. And again having worked at committees before there's a lot of incumbent Republican members of Congress who wish they had an 80% favorable rating with their base of Republican primary voters. So that just shows you what did rank choice voting do. And we'll get a little bit more into that in terms of why the campaign was positive.
So but that really kind of drove cow's favorability rating. And then when we look at not just favorability, but kind of the definition 23% of, of Republican primary voters. So we did the survey the day of the election after the polls closed and the day after. So all campaigning had been done. We did it right then. So you know, at that point voters recall was fresh, all that stuff. They didn't have the chance to kind of go back you know, especially in the 10th district, the nominee hadn't yet been announced. So even people who voted in the seventh district, 23% of them had either never heard of yes, Lee Vega or didn't have an opinion of her. For Cal in the 10th, only 6% had never heard of him or didn't have an opinion of him. And that's because with the rank choice voter ring choice voting voters have to be more informed.
So if you were a Lawson or Micone voter, you still needed to learn about hun cow and other candidates for your second, third and fourth choice. But in the seventh district crystal ook who was a supervisor, did supervisor, a lot of voters came out for her for in her geographic base. You know, they may be her neighbor. They, you know, know her from a supervisor. They really didn't need to know a lot about the other candidates because they were only voting for their one candidate. So again, it really makes the favorable ratings higher. It makes voters more informed about the candidates and really sets them up well for the general election, by having a very positive impression because you don't really have that negativity in a rank choice voting election where you're trying to build consensus.

Eric Wilson:

That's one of the things that proponents of rank choice voting like to point to is, are founding fathers of this country wanted and expected and informed electorate and, and rank choice voting where you're asked to have preferences on all of the candidates really does force you to, to, to learn about all of them, which is one of the reasons that we, we saw such high awareness and, and favorability for Hong cow in the 10th district. We also found that that voters in the rank choice voting primary and that 10th district thought the campaign was more positive than the voters in the traditional primary over in the seventh. I think it was striking that the difference was even more pronounced among moderate primary goers, walk us through those numbers and, and why a positive campaign is a benefit to a party.

David Kanevsky:

Sure. So at the beginning of the survey, we asked voters whether they thought the campaign for the Republican primary had been positive, negative, or a mix of both positive and negatives. And what we found was in both districts, people thought the campaign was positive, but it was much, much more positive in the temp district. In the seventh district, 50% of voters thought it was positive. 28% of voters said it was a mix of positive and negative. And 9% said it was negative in the 10th district. 84% said it was positive. 12% said it was a mix of positive and negative. And 2% said it was negative. So we're seeing that positive difference of about 34 points. So much more positive in the 10th district. Because again, when you have a very crowded field and you ever rank choice voting, it doesn't make sense to go after the front runner, because if they end up coming into a second or third place, you need their voters to kind of get over the hump.
We also saw, we asked intensity whether people thought it was a mostly positive campaign or somewhat positive campaign. Again, the intensity on kind of how positive it was also really shows up 24 points, much more positive in the 10th and in the seventh. And again, what I'd point out is, you know, positive campaigns, aren't a good in and of themselves, but what they really do is they help strengthen the candidates because cow had a much more positive image than Bega did as we talked about earlier. And why does that benefit Eric, your question about why does it benefit to not just the can, but to the party is if I'm hung cow now I don't have to spend as much time shoring up the base as Vega does. And that means cow and his campaign can now focus more on swing boaters.
and you really see that pronounced in as you mentioned with moderate voters in this case, what we would kind of define as softer Republicans and independents who voted in the primary. So in the 10th district, only 13% of voters of these moderate Republican voters said they saw either a negative campaign or a mix of positive and negative in the seventh district. The majority of these. So Republicans and independents saw negative campaigning of some kind. So again, 38 points higher in terms of kind of that negative level of campaigning. And the reason that's important is because a negative campaign, particularly with these. So Republicans, moderate Republicans can have an impact on the nominee heading into the general election. Maybe it pushes some G O P leaning voters off the nominee or decreases enthusiasm and turnout look at the Pennsylvania Senate, race odds one, you know, with just 31% of the vote. But if you look at a bipartisan AAR P poll among Republicans, he has a 53% favorability rating and 38% unfavorable. So you're seeing that the negatives from the primary are hurting odds in the general election.

Eric Wilson:

You're listening to the business of politics show. I'm speaking with David Kaneski about a recent poll. He conducted for the center for campaign innovation about rank choice voting. And we're talking about the, the impacts that this has on primaries. And I think one of the most surprising, I think our most surprising finding was that rank choice voting really benefited the runners up, which is, which is particularly powerful because you know, this year we've seen a lot of really competitive G P primaries where the Democrats have been much more effective at sort of clearing their fields. I'm curious to hear what benefits we saw for the, those runners up in the, the rank choice voting contest.

David Kanevsky:

So if you look at, in the 10th district, Jeanine Lawson came in second, she just got, you know, 29% on the first round of the boat. She ended up with 34% on the final round. Brandon Roone came in third. He he's a first time candidate started with just 10% on the first round 14% on the final round. But both of those candidates ended up with a higher net positive image than Nestle Vega did in the seventh district. So Lawson ended up with a net plus 59% favorable image. Micone ended up with a net plus 54% favorable image, both had very low unfavorability and they were both particularly strong in their geographic bases. Lawson was at 85% favorable in prince William Mishawn was at 74% favorable and loud. And so even though they didn't win, it sets them up well to run for office down the road because they've established a very positive image with Republican primary voters.
And quite frankly, if those candidates are listening or people who've worked with them are listening, they should go run for office again, please. You know, I know Jeanine Lawson is a supervisor. I don't know if she's gonna run for reelection or run for chairman of the board, but she's well positioned with Republican primary voters. If she wants to run for higher office, same thing with Ms. Shown. And again, they've established a good a good impression with the G O P base. So puts them well in a primary, but also lets them focus more on swing voters in the general

Eric Wilson:

Election. But that's not what we saw over in the seventh district.

David Kanevsky:

No, in fact, the initial front runner in the race, Bryce Reeves, he's a state Senator. He ended up with a much higher unfavorability ratings and in his state Senate district, there was a much higher level of negative campaigning, probably because strategically every campaign was thinking, well, I just need, you know, 30 to 35% of the vote. And if his district's about, you know, 25 to 30% of primary voters, it's like if he's, if he does really well there, that almost gets him enough to win. So a lot of campaigns were really attacking him, particularly in his state Senate district to try to hold his votes down. And so that will have some lingering impressions if he runs for reelection, it has implications even on Vega. Like she has some levels of negativity among Reeves and Anderson voters because you know, there may have been some contrast there that could have impacts in the general. So again, whereas, you know, not only do Cal and Lawson Macone, even Mike Clancy, Mike Clancy only got 5% of the vote, but had over 50% favorability rating. Those guys are well positioned to run for higher office, whether it's Congress, supervisor school board, because of rank choice voting and the positive campaigns we saw there. And

Eric Wilson:

It, it really speaks to the mechanics of rank choice voting, particularly in a primary where you have all of the incentives in a primary contest where it's the plurality, the single first pass the post in order to increase your vote, share you have to take it away from someone else. And so that's what we saw happening in the seventh with rank choice voting your path to victory is actually adding to your coalition. So it it's really interesting kind of game mechanic that's affecting the outcomes and how campaigns respond to 'em. We also learned a lot about voters' opinions of rank choice voting. You know, we, in the 10th district, they did this in 2020 for their nomination of Alicia Andrews. They participated in the statewide Republican party of Virginia chairman's race. They did it in the 2021 unassembled convention. And of course, again, in 2022, we found that 56% of voters in the 10th district preferred ranked choice voting, what do you think this means for, for advocates of ring choice voting?

David Kanevsky:

Look, there's no question. It's a positive finding. What we see is when voters are exposed to ring choice voting, they like it. I would say, as a pollster, you will always kind of look at both sides of an issue. The fact that a majority prefer it is good. We still have 40% who prefer the traditional method, particularly around older voters. And it's important for advocates of bring church voting to understand kind of why that is, you know, to be a conservative, especially a Berking conservative means, you know, you don't want change for changes sake, but change based in tradition. So that's why there's a natural, he sea among conservatives Republicans to kind of resist change, particularly a lot of push for rank choice voting has come in kind of in more liberal places like New York city, places like that. So what you really wanna focus on is you kind of walk through adoption at the local level and kind of build from the bottom up rather than kind of the top down, because once people are comfortable with it at the local level, it becomes a lot easier to adapt to higher up.
So, you know, if cities use it in city council, town, council, mayor elections, as I've seen them done in places like Utah or California, all of a sudden when once voters get exposed to it and they say, Hey, do we want rank choice voting for us? Congress, it'll become a lot easier to kind of get that adoption. And then the other part is also, this is kind of within the primary. So it's, you know, particularly with the Republicans it's within the family. So that makes it a lot easier of like, okay, I'm trying to decide. There's a lot of candidates. I like, which one do I want first or second? It's like, well, what flavor of ice cream do I want? I want all of 'em, but I have to choose. So let me rank for a second. And third, as opposed to like, you know, do I want ice cream or vegetables, you know in, in a general election.

Eric Wilson:

And one of the reasons that we looked at rank choice voting and measuring its effects in, in these two districts were because we we've seen over the last couple of decades. So between 2020 18, there were three the number of federal primaries with three or more candidates tripled. So this, this trend of crowded, competitive primaries isn't going away. You know, there are a lot of contributing factors here, whether that's social media, which tracks very closely with that timeline the citizens United reforms that, that allow wealthy individuals to push one candidate further, you know, all kinds of things at, at play where we're just not gonna see these competitive crowded primaries go away. And I'm curious to hear from your perspective, based on this, this survey data, how, how might reforms like rank choice voting be used to help Republicans stay competitive and, and win more races?

David Kanevsky:

The candidates who go through it end up with more positive images. So that helps Republicans in the generals because it unites the base and then allows them to pivot more to the general election. They don't have to heal the wounds from nasty primaries as a result. And then it also helps Republicans, not in the short term from just, Hey, win this election. But long term build a bunch of candidates because even the candidates who don't win end up with positive impressions among voters that lets them run for future offices. So, and quite frankly, it probably helps those candidates want to run again because they're not, they're not dealing with the negativity of campaigns that say, you know what, I don't wanna give this another shot. I don't want my neighbors, my kids to hear all this nasty stuff. So so those are the things that helps Republican candidates win elections and helps recruit more Republican candidates to run for office.

Eric Wilson:

All right, Dave, we ask everyone this question. Is there a, a startup that you'd like to see, maybe there's a problem or challenge out there that you think's unaddressed that would make for a good product

David Kanevsky:

In politics. We use consumer data to improve a lot of our political targeting. The question is how can people use political data to market, to consumers? And we do see some of this on the right, you know, there's films clearly aimed at Christian audiences. I, I did some races in Louisiana, several cycles ago, duck dynasty, their endorsement was very popular, things like that. And we do see entertainment shows on the left. That very much look like they're trying to find actively liberal audiences like in their messaging, like the daily show or SNL. So I would argue for what would be a production company that focuses on making TV shows and movies that I wouldn't argue are political, but aimed culturally like hardly conservatives, you know, that production company may not win a lot of EES, but as we see a very fractured media environment, could that be a, a financial success by holding an attract an audience? You know, what would sex in the city or modern family look like? That's aimed at that audience? An analogy I point to that is look at coin the new credit card company that Cory Gardner's involved of. They're appealing to conservatives by offering a percentage of the profits back. So I would argue that that that is something like could make some money, maybe have some like very downstream cultural impact, very politically driven, but just kind of aimed at, you know, Heartland conservative values could, you know, be an interesting business position.

Eric Wilson:

Well, my thanks to David for joining us today, you can learn more about this poll@campaigninnovation.org. There's a link in the show notes. If you wanna learn more about what we found and, and dig into that data, please remember to subscribe to the business of politics show wherever you listen to podcasts. And if today's episode made you just a little bit smarter or gave you a, a new perspective on something, please share it with a friend. It helps us get it in front of more people and it makes you look really smart. So please do that. And if you've been listening for a while, remember to leave us a rating and a review that really helps with reaching more people so we can get out our message of bringing innovation into politics and better understanding the business of politics with that. We'll talk to you next time. Thanks for listening.

‍