Should We Believe the Latest Iowa Caucus Polls?

If we are to believe the polls, Republican frontrunner Donald Trump appears to be breaking away from Texas Sen. Ted Cruz with 4 days to go before the Iowa caucus. But can we believe them? Historically speaking, primaries and caucuses are notoriously difficult to poll because, setting aside the difficulty of forecasting electorate composition, they tend to break late as soft and undecided support shifts unpredictably.

A look at polling in the two previous cycles tells the tale. First, because PPD didn’t aggregate polls until 2014, let’s take a look at the RCP average of Iowa caucus polls juxtaposed to actual caucus results.

2008 Iowa Results

1^st – Huckabee 34.4

2^nd – Romney 25.2

3^rd – Thompson 13.4

4^th – McCain 13.0

5^th – Paul 9.9

The reason I started with 2008, when Arkansas Gov. Mike Huckabee won the caucus, was to drive home several points. First, sometimes trends don’t even allow us to gauge how these things are going to break. Huckabee was actually on the decline and Romney, who was leading going into Tuesday, was on the upswing. Of course, he ended up underperforming in the caucus by 3 points and the winner underperformed in the polls by 6.8.

Still, from 3rd place to further at the end of the pack it gets a little hairy, but eventual nominee John McCain performed better than his poll position yet still fell to fourth place behind the late-great Fred Thompson. Let’s take a look at the final Iowa caucus polls during most recent cycle, as well as the final results.

2012 Iowa Results

1^st – Santorum 24.6

2^nd – Romney 24.5

3^rd – Paul 21.4

4^th – Gingrich 13.3

5^th – Perry 10.3

The 2012 Iowa caucus polls were even more off the mark, a la at least they picked the winner correctly in 2008. A distant fourth place in the polls, most pundits were saying if only Rick Santorum had another week before the caucuses he would take. Well, he did take it by outperforming by 14.6 points, while Romney, who was first in the polls, outperformed by just 3.1 points. Ron Paul’s vote share was actually pretty close to his end result, but it obviously wasn’t enough.

The point has been made. The Iowa caucus has been a notoriously difficult contest to poll in recent cycles and, with Trump and Cruz bludgeoning each other, the potential for someone else like Marco Rubio to outperform is real. Though there is no real comparable example on the GOP side, it happened to Howard Dean in 2004 on the Democratic side.

That said, according to the PPD average of aggregate Iowa Caucus Polls, Trump has a substantially larger lead than other candidates in the last two previous cycles, currently 6 points. Further, with the exception of only a few short weeks, he has maintained that lead far longer than other candidates. Worth noting, we have spoken with election officials on the ground and they tell us there is a tremendous increase in caucus activity and requests by independents to register by party, a requirement to participate in the caucus.

Of those new registrations, one officials told PPD “they are overwhelmingly going on the Republican side.”