DMR: escenarios

Des Moines Register:

At some point, we have to stop interviewing for our Iowa Poll and publish our findings. That is just a fact of life.

But what if there is a late-breaking surge of evangelicals who decide to caucus at the last moment? What if seniors, who appear less likely to attend the Republican caucuses this time than in 2008, change their minds and come out caucus night in numbers similar to last time? What if disaffected supporters of Barack Obama decide to caucus on the Republican side this time around?

Yogi Berra famously said, “It’s tough to make predictions, especially about the future.” Well said. We can, however, use statistics to test what could happen under these scenarios. Here’s what we found when we did that:

&#; EVANGELICALS — In 2008, they were 60 percent of those who participated in the entrance poll at the Republican caucus and famously handed Mike Huckabee the win. This year, our polls are showing far fewer likely Republican caucusgoers identifying themselves as born-again or evangelical Christians — about one in three.

When we weight our data from the last two days in the field to match 2008, Rick Santorum wins with 25 percent, Mitt Romney is second at 20 percent, and Ron Paul has 16 percent. Michele Bachmann, Newt Gingrich and Rick Perry barely break into double digits. Each is within a point of one another.

&#; SENIORS — In 2008, they made up 27 percent of the entrance poll respondent pool. We are showing them under 20 percent in our polls this year. When we weight our data from the final two days of polling to match 2008, this greatly benefits Romney, who would rise to 26 percent and a 7 point lead over Santorum, 19 percent. Paul drops to third place, with 17 percent.

&#; INDEPENDENTS — They were just 13 percent of the entrance poll sample in 2008. In the last two days of interviewing, we saw elevated numbers of people who self-report their party affiliation as independent (26 percent).

What if there is a late surge of independents? We tested weighting independents to be an unlikely 30 percent of likely caucusgoers. That narrows the spread, with Romney holding a 2 percentage-point lead over Paul, followed by Santorum.

&#; DEMOCRATS — Our poll methodology relies on sampling only from registered Republicans and independents. We’ve elected to exclude registered Democrats because of the cost of interviewing the small numbers of Democrats likely to attend a Republican caucus, with little expected influence on the final outcome.