Our final CBS News Iowa Battleground Tracker offers a statistical simulation of the caucuses and some scenarios that might unfold on Monday. It looks like a close contest heading in, and the top candidates are all poised to win national delegates.
To show what could happen — and more importantly, why — we continued interviewing likely caucus-goers this week for their first- and second-choice preferences in our polling, then combined it with data on Iowa voters generally, and how the caucus system works across the state's counties and districts.
Monday dawns with Bernie Sanders and Joe Biden even in first-choice support at 25% each in our baseline model, Pete Buttigieg very close behind at 21%, and Elizabeth Warren at 16%, also in position to accrue some national delegates. Amy Klobuchar is at 5% in our baseline estimate, and all other candidates are under 5%.
Now the fun starts, as we look at what could unfold both in terms of caucus-goers' final preferences and turnout.
Some of Biden's core strengths emerge in our "How Biden Can Win" scenario. He's the second choice for many supporters of lower-tier candidates, who would have the opportunity to switch support during the caucus meeting if their
Biden is the most likely second choice among Amy Klobuchar's backers, in particular. In a Biden-win scenario, they and others below the viability line would shift to him, pushing him slightly over the top.
Biden's other advantage is that his support is spread out fairly evenly across the state. To win the Iowa caucuses, a candidate essentially has to win delegates from every part of the state. Biden's support would add up piece-by-piece to a narrow statewide edge for him.
In that case, Biden would get to 27% and 13 national delegates out of Iowa; Sanders, with 26%, would receive 10 delegates; Buttigieg, at 23%, would receive 10 delegates; and Warren, 18% and 8 delegates. But we should also note that Biden's vulnerabilities include having supporters who haven't been enthusiastic about him and have been considering other candidates in our recent polling. That could open the door for Sanders, among others.
Sanders has a couple of paths underpinning our "How Sanders Can Win" scenario. First, in contrast to Biden, Sanders shows a core strength in that his backers have been the most enthusiastic of any top candidate's. If that translates into a turnout boost, as recent polling suggests it could, then Sanders would gain in the vote count. However, Sanders' challenge is that his support appears concentrated in urban areas, in part because of backing from young people. A turnout surge in the largest counties could propel Sanders to a lead in the statewide preference tallies. However, Sanders could rack up even more delegates if this turnout increase happened more evenly across the state.
Second, Sanders could get an additional boost from Warren backers, if she were to fall below the 15% viability threshold in some caucuses, as a plurality of her supporters pick Sanders second. We estimate that Warren is just above the threshold statewide, so she will probably be viable in at least some of the caucuses. If she were to become unviable in the counties where her support is weakest — and notwithstanding other potential movements — Sanders would benefit from the realignment: His support would increase to 26% and 13 delegates, while Warren's would shrink to 14% and 7 delegates. Biden would end up second at 25% and 11 delegates; and Buttigieg 21% and 10 delegates.
Buttigieg appears to need strong turnout from first-time caucus-goers, as more of his backers are prospective first-timers than supporters of other top-tier candidates. That could certainly happen, as more than 4 in 10 people have been first-timers in recent caucuses, according to CBS News entrance polls. He could also gain if supporters of lower-tier candidates disproportionately shift to him over Biden, as Buttigieg and Biden are close in these voters' second-choice support.
And Warren is above the cutoff line at 16% statewide —— and in position to pick up national delegates. She stands to gain if Sanders' backing isn't spread out across the state. She could also gain in places where the three candidates ahead of her split the vote more evenly. In particular, Warren could gain in places where Klobuchar's supporters do not make threshold. Besides Biden, Warren (and Buttigieg) are their second choices more so than Sanders.
And one final reminder: Caucuses are meetings where people can and do change. After all these months we know a lot now about who voters like and who they're considering, and why. But viewers and political junkies also know that deliberation and persuasion are critical parts of what goes on, and what makes watching the results so exciting on Monday night.
These Battleground Tracker delegate and vote preference estimates in Iowa are derived from a multilevel regression model fit to survey data collected between January 22-31, 2020. The model incorporates individual-level and district-level variables, such as age, race, gender, education, party registration, and vote history. Vote intentions were imputed onto voter file records in Iowa and then aggregated statewide and by district. The margin of error on these estimates is ±3 points. It reflects the range of uncertainty around the estimates produced from the model simulation.
The model partially pools data across states including Iowa, estimates effects for each state and district in the sample, as well as a linear time trend. Estimation of voting intentions is based on surveys conducted for the Battleground Tracker project, including a sample of 1,835 registered voters in Iowa and a nationally representative sample of 34,375 registered voters from YouGov.
Of respondents who participated in the previous wave of Iowa polling from January 16-23, 2020, 538 were re-interviewed for this study — a recontact rate of 29%. Respondents were selected to be representative of registered voters within Iowa. 732 were selected from YouGov's panel, and 1,103 from other online panels (Critical Mix, Dynata, Lucid and Precision Sample). All candidates were shown to respondents to express their first and second choices.
Using estimated vote intentions, we simulate the number of delegates going to each candidate by district and statewide. Following party allocation rules, candidates with less than 15 percent of the vote in a district or at large are eliminated, and delegates are awarded to remaining candidates in proportion to their vote shares.
Delegate estimates indicate how many delegates each candidate would receive based upon preferences respondents currently express. They are offered to explain how vote preferences translate to delegates under party rules and are not a forecast — the model does not incorporate forward-looking uncertainty. Read more about the process and the Battleground Tracker.