Vermont Sen. Bernie Sanders’ odds of winning the Iowa caucuses have been rising over the last month. Even the worst polls for Sanders have him close to the top. Yet some data suggests some potential weaknesses in Sanders’ armor, given how caucuses differ from primaries.
A new Monmouth University poll, for example, has Sanders (21%) well within the margin of error against former Vice President Joe Biden (23%). The previous Monmouth poll had put Biden at 24% and Sanders at 18%.
When you add the Monmouth poll to the average of high-quality recent polls, Sanders comes in with 22% to Biden’s 18%. Not too far behind are former South Bend, Indiana, Mayor Pete Buttigieg at 17% and Massachusetts’ Sen. Elizabeth Warren at 16% in the average.
Given the predictiveness of past Iowa polling at this point, candidates in Sanders’ position win about 3 in 8 times. Biden is not too far away, with about 2 in 8 times. Buttigieg and Warren round them out with each winning between 1 in 8 and 2 in 8 times. Compared with where we were over the weekend, Sanders is up. Buttigieg and Warren, meanwhile, have seen their chances shrink. Biden is pretty steady.
You’ll notice, though, that even if Sanders is the favorite, he isn’t a heavy favorite. Iowa polls haven’t been all that predictive of the results. If you were putting a 95% confidence interval around the average given past polling predictiveness, it would be +/- 16 points around the top candidates. Indeed, it’s especially difficult to game out the caucuses on the Democratic side.
Remember, there will be an initial ballot test at each caucus site. Any candidates who can’t reach at least 15% (depending on the caucus site) on the first ballot will not be considered “viable,” making them ineligible for any state delegate equivalents at said site. These candidates’ supporters then have the chance to realign with a viable candidate.
It’s therefore key that candidates remain liked by supporters of the other candidates who won’t reach viability.
The Monmouth poll suggests Sanders may be losing some popularity outside of his base over the last few weeks. His net favorability rating (favorable minus unfavorable) dropped from +48 points earlier this month to +29 points now, which is a dramatic slide. No other candidate had a decline of more than 5 points.
Additionally, Monmouth asked likely caucusgoers who they would support among the top four candidates (Biden, Buttigieg, Sanders and Warren). In its last poll, it was Biden at 28% to Sanders’ 24% on this question. This time it was Biden at 29% to Sanders’ 25%. In other words, there was no movement in Sanders’ direction, even though he picked up ground on Biden in the initial preference test.
Now, it’s possible that Monmouth will end up being an outlier in this regard. But it’s also possible that the foundation beneath Sanders’ feet isn’t as solid as it may seem. Things can change really quickly in Iowa.
Back in 2012, for example, Republican Rick Santorum, the winner that year, was gaining 3 to 4 points per day during the final week before the caucuses.
A candidate like Sanders with declining popularity could conceivably lose a few points in the final few days. That could make all the difference in the world in the final results.
