Smaller, Better — And Weirder? How We'd Change Who Makes The GOP Debate Stage.
![](https://fivethirtyeight.com/wp-content/uploads/2023/08/538_WhatIfDebates_Example2_4x5_v01_ag.png?w=859)
ABC News Photo illustration
Welcome to (a version of) FiveThirtyEight’s politics chat. The transcript below has been lightly edited.
maya (Maya Sweedler, senior editor): Well, my friends, the Republican National Committee’s debate qualification period has officially come to a close, meaning the first Republican primary debate is set. We’re confident that nine candidates have qualified, via polls and donors, to be on stage in Milwaukee on Wednesday: former President Donald Trump, Florida Gov. Ron DeSantis, entrepreneur Vivek Ramaswamy, former Vice President Mike Pence, former U.N. Ambassador Nikki Haley, former New Jersey Gov. Chris Christie, Sen. Tim Scott, North Dakota Gov. Doug Burgum and former Arkansas Gov. Asa Hutchinson. Now, there are a few other candidates who argue they are qualified, but as of last night, we don’t have RNC confirmation. And of those nine, not all have completed the final step to guarantee their spot — signing the RNC’s loyalty pledge — so we can’t be sure all will show up. (In fact, we’re pretty much positive at least one will not, as Trump has said that he does not intend to sign the pledge.)
But let’s back up. When announced, the RNC’s debate-qualifying requirements seemed pretty similar to the Democrats’ requirements four years ago — just a little more stringent. The RNC required candidates either earn 1 percent support in three national polls or at least 1 percent in two national polls and two polls from the first four states voting in the GOP primary (each coming from separate states), based on surveys conducted by organizations unaffiliated with a candidate that sampled at least 800 likely Republican voters. Candidates also had to earn donations from at least 40,000 unique contributors, with at least 200 donors from 20 or more states and/or territories.
These criteria didn’t satisfy every candidate. As recently as last week, Burgum — who qualified for the debate after running a gambit to attract donors by offering a $20 gift card for $1 donations — complained to Politico about the RNC’s debate qualifications, saying, “If you have clubhouse rules, then you design the rules; they always favor someone, and they always disfavor others.”
Well … yes. But how wacky can these clubhouse-favoring stages get? After staring at the actual RNC rules these last few months, some of us decided we wanted to design them in order to create different — maybe better? — debate stages. So we went back to first principles, decided what we wanted our stages to do and then built them out using RNC-like criteria.
gelliottmorris (G. Elliott Morris, editorial director of data analytics): OK, so the big question for me is about optimization: What are the primary rules trying to actually solve?
I’m a big fan of the political science literature that says uber-democratized primaries are, somewhat paradoxically, bad for democracy, and especially bad for candidate selection and optimizing representation (because of various types of bias: e.g., interest group capture, potential polarization from ideologically fringe participants, confusion from too many options, etc.). So I would want new rules to work toward solutions for those things, rather than just limiting the number of people who appear on the stage
If I were the RNC, I would enact a new rule that taps into the quality dimension of the candidates, rather than just the popular support dimension. I would require each candidate to:
- Earn at least 1 percent in an average of the latest national poll from each pollster publishing a survey in the 30 days before the debate. The population of the poll is all adults who affiliate with that party.
- In at least one of the first four primary states, earn at least 1 percent in an average of the latest poll from each pollster publishing a survey in the 30 days before the debate. (Same population restriction as above).
- Be endorsed by some number (five? Ten?) members of Congress (perhaps with the additional restriction of them being from outside their home state).
Finally, if these criteria yield more than six qualified debaters, you rank them by the percentage of partisans who say they would consider voting for that candidate, and split them up into two debate nights. If there are more than 12 qualified candidates, those in the 13+ positions don’t get to debate.
Based on our endorsement tracker and Geoffrey’s excellent tracking of candidates’ donations and polling thresholds, the debate stage would only be Trump and DeSantis.
![](https://fivethirtyeight.com/wp-content/uploads/2023/08/debate-stage-morris.png)
geoffrey.skelley (Geoffrey Skelley, senior elections analyst): For context, a head-to-head debate this early in a primary cycle would be unheard of. In fact, there’s never been a single pre-Labor Day presidential primary debate in the year before a primary featuring just two candidates, based on debates from the past couple of cycles and an old list I helped put together in 2015.
gelliottmorris: Readers may think that’s an awfully small debate stage, but remember that candidates so far have not been required to woo congresspeople — and, equally, congresspeople have not felt it was their responsibility to shape the outcome of the primary. If this were an incentive for the primary, I think several other candidates would have qualified under these rules by now, including Pence, Scott and Burgum (who are on the cusp of the congressional endorsement requirement already).
maya: Amazing! A two-man debate stage during the primary. Your rules, Elliott, also pretty much ensure that participants will hold major federal or state office. (Sorry to Francis Suarez and all the mayors who have come before him.)
One thing I think makes this stage particularly interesting, especially for Wednesday, is that DeSantis appears to have decided that aligning himself with Trump — and against the other Trumpy candidates — is a winning strategy. In a series of memos about his debate strategy released online, DeSantis’s team stressed the need to “defend Trump” and “hammer” Vivek Ramaswamy, who’s striking pretty Trumpy notes as he climbs ever so slightly in the national polls.
So without the distractions of anti-Trump voices, it could be a fascinating triangulation for DeSantis, who has really struggled to carve out support from the Trumpy lane of the party but whose rhetoric and policies are not going to fly with the less Trumpy part of the primary electorate (to say nothing of the general …).
gelliottmorris: Yeah, uhh, about that memo … Readers should think of the GOP primary as actually being two primaries right now. One is a primary between Trump and a hypothetical alternative, and one is a primary to be the alternative. The problem with the DeSantis strategy is that even if it helps him in the minor league primary it hurts him against Trump, who is the ultimate opponent here.
A smaller debate stage means voters would get a better debate about the biggest dimension of conflict in the GOP this year: Whether the party should move beyond Trump to better debate the issues that separate them from President Biden.
geoffrey.skelley: DeSantis could benefit from a lot of attention in this debate. And defending Trump is understandable considering how well-liked Trump is among Republicans. It does present challenges, however, for DeSantis if he wants to knit together a rickety coalition of primary voters who like Trump but want to move on with someone else, pro-Trump voters who he might win over and voters with an unfavorable view of Trump.
Obviously, DeSantis would relish a one-on-one matchup with Trump in a debate, under Elliott’s rules.
gelliottmorris: I guess the ultimate question for me is just … What do people want from this? And what are we getting out of it? If the RNC wants to highlight the “loyalty to Trump above all else” dimension of Republican/”conservative” ideology today, then a 10-person debate works in their favor — but that almost certainly hurts them in the general election. A contest of ideas would likely better serve both Republican voters and the RNC.
DeSantis vs. Trump — that is a contest both of ideas and alternatives. Trump vs. Christie works too, but much less on policy. I struggle to think of what the primary would be if it were just Trump and someone else instead. Take Trump vs. Ramaswamy. What would the conflict even be?
maya: I have a sneaking suspicion that a Trump vs. Ramaswamy debate would turn into the equivalent of watching two buddies go fishing.
geoffrey.skelley: All right, so for my debate stage, I would want to focus on three things: polls, donors and favorability. I think that the RNC’s polling and donor requirements have largely been good — although I wish the RNC would publicize what polls can count toward qualification. But as a means of limiting one pollster’s influence, I would do what the Democrats did in the 2020 cycle and permit a candidate to count only one poll from a pollster in a given geography. That way, you can’t rely on, say, three national surveys from one pollster who has found better numbers from you to make it. I’d also maybe rule out any pollster that used interactive voice response in its polling, just to avoid some potential junk.
But I also think the debate stage should only have individuals whom Republicans actually like. So while my initial group of candidates would look similar to the eight who had enough polls and donors to qualify as of late last week, I would also look at favorability polling and use that to cut the list down by only including candidates who have a net positive favorability rating.
For instance, Christie has a miserable favorability rating among Republicans overall, so he would be out. Burgum remains largely unknown and is actually net negative in the few polls that have asked about his favorability. As a result, the real question would be whether Pence is slightly positive or negative when the poll arrived. At this moment, he’s just barely in negative territory, so if he were excluded, we’d have the following five debaters: Trump, DeSantis, Ramaswamy, Haley and Scott.
![](https://fivethirtyeight.com/wp-content/uploads/2023/08/debate-stage-skelley.png)
I believe that is a group of candidates that Republicans would actually clamor to hear from, and it would present opportunities for DeSantis, Haley and Scott (maybe not so much Ramaswamy) to differentiate themselves from Trump.
maya: Ooh, interesting! What’s your bone to pick with IVR?
geoffrey.skelley: That was just an example. I guess if it were a part of mix-modes, it’s not the worst thing. But one of the downsides to the RNC’s approach is that you can have pollsters that rely heavily on modes like IVR or online panels rack up large sample sizes that may not actually be that great quality-wise, even if they meet the 800+ sample size the RNC wants. However, I get why the RNC didn’t want to use the DNC’s approach from 2020, because the DNC essentially excluded some pollsters ahead of time, which could prompt candidate complaints if they did well in a poll that didn’t count even if it was from a respected pollster.
gelliottmorris: Geoffrey, I don’t love IVR either, but I wouldn’t want to toss it out entirely if we didn’t have to. You know, there is a data journalism website out there that controls both for mode of interview and pollster frequency (your other concern) … The RNC could just use our numbers!
#FiveyForPresident
maya: ÑÐÐÑÐѤÐÐâ¬
geoffrey.skelley: Haha, this is true. I think the pollster frequency is my main complaint. For instance, former Arkansas Gov. Asa Hutchinson would have only narrowly had enough national polling if he’d only been able to count one national poll each from Morning Consult, JMC Analytics and Kaplan Strategies, which collectively gave him eight qualifying polls (versus three under my rules). Granted, he’d still have had enough early-state polls to make it, but it’s one example.
Whatever happens, I guess the 2028 cycle might see one or both parties smartly use the best practices from the 2020 and 2024 cycles for determining debate qualification. Something to look forward to!
(Since they are unlikely to do what Elliott proposes by more aggressively limiting the number of participants.)
maya: I am going to go the total opposite direction of smart.
I want my debate stage to introduce viewers to as many lesser-known candidates as possible. Candidates who have spent years describing their vision for America from various national platforms don’t need 45 seconds between ripostes to present their ideas; candidates who, because they represent smaller states or are newer to politics or have been out of the game for a while now, would benefit more from this opportunity.
So in the interest of exposing primary voters to as many new faces as possible, I propose an elimination-based debate guideline system. I would take the entire pool of candidates and eliminate the following:
- Candidates whose national name recognition (which I’m defining as the sum of a candidate’s favorable and unfavorable ratings among all adults — because hey, a national audience is a national audience) is in excess of 65 percent.
- Candidates who have polled in the double digits (limited to polls of adults of the candidates’ party) nationally at least three times in the debate-qualifying period.
- And, just to get a little weird, candidates who at any point have earned more than 1 million votes in any election. (And we’re including all members of a ticket on this — sorry, Mike Pence, but you’re out three times over based on this.)
Interestingly, this last requirement would eliminate Scott, who earned just over a million votes in South Carolina in his 2022 reelection bid, but not Nikki Haley, who topped out around 700,000 votes in her 2014 reelection as governor of the same state. Haley also squeaks under the name recognition limit.
Given these rules, my stage would be: Haley, Hutchinson, Burgum, Suarez and Hurd.
![](https://fivethirtyeight.com/wp-content/uploads/2023/08/debate-stage-sweedler.png)
Ramaswamy would be eliminated by the double-digit polling standard, and Christie by the million-vote threshold (the man got more than 60 percent of the vote in his 2013 reelection bid in New Jersey).
gelliottmorris: Woah! Wild card rules!
geoffrey.skelley: Is this like a full-on “undercard” debate — see: the GOP’s 2016 primary debates — as the main event?
maya: Let’s get WEIRD.
Yeah, Geoff, exactly. Who needs to see Trump and DeSantis go back and forth again when both have spent hours (literally, HOURS) on national TV already?
gelliottmorris: I think this is neat if you put the debate in the context of the wider primary. With these rules, debates effectively become free advertising for lesser-known candidates (who might not be able to afford endless expensive TV ads) and decreases the weight we put on The Discourse about candidates who voters already know a lot about.
But I would still want to hear from everyone on the same stage at some point. Maya, maybe we could do this for the early debates and use a more traditional screener in, say, December and January, once the name recognition playing field has been equalized? Or maybe I need to embrace the weird!
geoffrey.skelley: Maya really wants a Jimmy Carter-style candidate to be possible again, but in the sense of gathering more national exposure ahead of the Iowa caucuses rather than largely via them.
maya: Exactly!
geoffrey.skelley: Well, Maya’s approach might ensure we have 35 presidential candidates next time around, so maybe we can have a couple of “undercard” debates to kick things off in 2027.
gelliottmorris: Thirty-five candidates! We’re going to have to put them in The Big House!