The Gallup organization released their likely voter model on Oct. 26, which, along with Rasmussen Reports, has consistently shown much better numbers for Mitt Romney than other national polls. It is important to note that Gallup and Rasmussen are the only two pollsters that have tracked the presidential race every day for the entire election cycle. Both have found consistently similar results, too. The myriad of other pollsters have conducted benchmark or brush fire polls, meaning they are surveying the national electorate for just a short period in time and producing a snapshot of the voters’ intentions rather than a trend.
Right now, all of the polling suggests a Romney lead in the national popular vote among those considering themselves likely to vote, with Pres. Barack Obama doing better in the critical states and among the registered voter universe. The likely voter numbers are producing a very unique and inverted situation because similar situations in the past have always shown the Democratic candidate leading the popular vote, while the Republican fared better in the states.
According to the Gallup analysis, the electorate looks virtually the same as it did in 2008, but the voting intensity model is quite different. There is either no, or only a one-point, difference in the demographic categories when comparing today with four years ago, and as much as a four-point increase among non-white voters when overlaying 2012 data against what was found within the 2004 electorate that re-elected George W. Bush.
Gallup maintains that their likely voter model, culled from a series of seven screening questions (see below), has correctly predicted the final trend in the past two elections. Their 2004 pre-election projection suggested a two-point Bush popular vote win, which is what happened. In 2008, the final data correctly predicted Obama’s win, but over-shot his performance. The Gallup data predicted an 11-point Obama spread; in actuality the final count was +7 points over Republican John McCain.
Today’s model shows virtually no demographic difference between 2008 and 2012, but a major swing in terms of self-identified party registration. Four years ago, 39 percent of the likely voter sample considered themselves Democrats to only 29 percent for Republicans.
According to Gallup’s latest data, the 2012 partisan self-identified likely voter ratio breaks Republican 36-35 percent, a swing of 11 net points for the GOP (+7R; -4D) in comparison to 2008. When the “lean Democrats” and “lean Republicans” are added, the model expands to 49-46 percent Republican (based upon tracking data collected over the Oct. 1-24 period). This is highly significant in detecting electoral intensity. If correct, the 2012 vote will be very different from 2008 and much more Republican. Today, Gallup shows a 50-46 percent spread in Romney’s favor among likely voters. Rasmussen finds Romney’s lead to be a similar 49-47 percent.
But, it all comes down to which of the pollsters’ likely voter model is correct. Gallup has actually posted the seven questions they ask to determine voter participation intent, as reported on Gallup.com.
- Thought given to election (quite a lot, some)
- Know where people in neighborhood go to vote (yes)
- Voted in election precinct before (yes)
- How often vote (always, nearly always)
- Plan to vote in 2012 election (yes)
- Likelihood of voting on a 10-point scale (7-10)
- Voted in last presidential election (yes)
Answers are graded on a scale of 1-7 and the results categorized accordingly. The latest numbers from their registered voter pool gives Obama a 48-47 percent edge, but the likely voter group goes significantly for Romney, as previously mentioned, 50-46 percent.
It’s going to be a very close and interesting election. Next week will determine which of all predictions are correct, but Gallup has already provided the most information to help us understand their support methodology.