You may not have heard in the various post-mortems of the Iowa Caucuses, but the public polls got the caucus results pretty much right.
Now, “pollsters get it right” isn’t exactly a juicy story, so we can all forgive the media coverage for moving on to more interesting things such as “Pompous Baloney”, but I’d like to take a moment to give credit where it is due for the good work that the public pollsters did in Iowa.
Below is the average performance of the candidates across the post-Christmas public polls along with the actual Caucus results:
|
Average Post Christmas |
Caucus Result |
Difference |
|
| Romney |
21% |
25% |
+4% |
| Santorum |
15% |
25% |
+10% |
| Paul |
20% |
21% |
+1% |
| Gingrich |
14% |
13% |
-1% |
| Perry |
11% |
10% |
-1% |
| Bachmann |
8% |
5% |
-3% |
| Huntsman |
3% |
1% |
-2% |
| Undecided/Other |
8% |
- |
|
Other than two changes that any cogent analysis of the polling predicted, the average of the post-Christmas polls was pretty much spot-on. The rise of Santorum as the conservative non-Romney alternative was both predictable and clear in the trends. It should have surprised few people that he took the lion’s share of the remaining undecideds. Similarly, the handful of voters who finally “settled” for Romney as the frontrunner is unsurprising.
While giving credit to the polls for their work in Iowa, it is important to keep in mind that Caucuses have some features that make them relatively easy to poll:
- The turnout is relatively stable across competitive caucuses. This makes it easy to build a likely voter model and sample. This year’s “record” Caucus turnout was around 122,000, up only slightly from the 119,000 participants in 2008.
- Though an unaffiliated voter could technically become a Republican to vote in the caucuses, even on caucus day, few do. This makes it easier to identify eligible voters and limits the complex issues of cross-over voting with which a pollster must deal.
- The Caucus process is relatively complex. It requires a greater time commitment than normal voting and must be attended at a fixed time. This higher barrier to participation reduces the number of “casual” voters and again simplifies the pollster’s job.
If Iowa is an easy task, New Hampshire is difficult. It is much easier for an Independent to participate and New Hampshire Independents have a history of doing so which causes turnout to vary considerably. It is also a more standard voting structure than a Caucus, making it easier for voters to participate.
The challenges that pollsters had in New Hampshire in 2008 have been well documented. But the 2000 primaries had their own problems.
So what should the observer be watching for tonight:
- Turnout. We’ve been hearing (and seeing in the polls) the story of a primary electorate not enthusiastic about any of the candidates. Does that translate into a lower-than expected turnout in New Hampshire? If so, does that benefit a Ron Paul? Or maybe a Jon Huntsman?
- Independents. In something the reverse of the above turnout story, Independents in New Hampshire only have one competitive Presidential primary race in which to vote this year. If all of those who habitually participate vote in the Republican primary, does this give Mitt Romney (who has been publicly tacking toward the center in New Hampshire) a further bump?
- The late bump. After 2000 and 2008 pretty much every pollster studying the issue has concluded that a disproportionate number of New Hampshire voters seem to change or make up their minds late. If there is a late break this year, it would seem to benefit a candidate like Jon Huntsman who has been trending upward in the state or perhaps Iowa runner-up Rick Santorum who saw a bump after Iowa (though that seems to have stalled).
Regardless of the outcome tonight, we’ll either be treated to a new round of stories about “how the polls got it wrong” or, like Iowa, no one will notice them having gotten it right.
