Mitt Romney, Frontrunner

Bill Dalbec is senior vice president of APCO Insight^®, the opinion research group at APCO Worldwide.

The results are in from New Hampshire, and there’s a positive story for organizations that polled until the end. New Hampshire is a tough state to poll in presidential primary elections because independents can choose to vote in either primary, and residents can choose to register on-site the day of the primary and participate. In addition, there is no single poll that has consistently been able to crack the code and provide spot-on results prior to the vote. As a result, the survey results can be all over the place and not very insightful or predictive. This is the opposite of what we typically see in Iowa, as I discussed last week. With an open primary coming up in South Carolina, what are some lessons we can take from New Hampshire to make us better informed poll readers and pundits?

The New Hampshire results again highlight the importance of survey interviewing timing in primary elections. Three polls were published and received widespread attention on the eve of the New Hampshire primary. Two of the polls interviewed through Monday night (Suffolk University/7 News and American Research Group), and one interviewed through Sunday (WMUR/University of New Hampshire). In addition, the two latest polls had two-day samples, whereas the WMUR/UNH survey had a four-day sample. In the end, the Suffolk/7 News and ARG polls both caught Jon Huntsman’s late surge, but the WMUR/UNH survey missed it. Polling until the eve of the election with a two-day sample enabled these two surveys to pick up on Huntsman’s movement at the end (the exit poll indicates most of his voters decided to cast ballots for him in the final days before the election).

Sampling is still critical. An open primary causes all sorts of havoc as the pollster tries to guess who will show up and the correct proportions. While it was generally accepted that Mitt Romney would win by a large margin and that the race was for second place, none of the three polls correctly identified that Ron Paul would secure second place the way he did. One explanation may be that the pre-election polls underestimated how many younger voters would turn out. The exit poll estimates that 18- to 29-year-olds comprised 12 percent of electorate; those polls that published demographic figures had only about 6 percent of their samples comprised of these younger voters. Given that Paul won this age group 46 percent to 26 percent over Romney and that one in four of his voters fell into this age group, that’s why the pre-election polls missed his clear second place finish.

Mitt Romney is now the de facto incumbent. This means that what Romney gets in the final poll is likely what he’ll get on Election Day (within the margin of error, of course). We now have two examples of this in Iowa and New Hampshire: the final Iowa Poll had Romney at 24 percent and he received 25 percent of the caucus votes, while in New Hampshire two polls pegged Romney’s support at 37 percent and he took home 39 percent. Late deciders will probably not be breaking to Romney; he is the person they know, and most voters who are undecided at that stage are looking for an alternative. The New Hampshire exit poll shows that Romney got 32 percent of the votes of people who made up their minds in the last few days before the primary but captured 49 percent of the vote among those people who decided more than a week before. A similar pattern was evident in Iowa.

Romney is currently around 30 percent in the South Carolina polls (which are all over the map thanks to the open primary). For those paying attention to the polls, don’t be surprised if what Romney gets in the final published quality poll is what Romney will likely get in the Saturday election.

Stay tuned.