By now, many of you can probably recite my near-weekly presidential polling sample admonition by heart. For the uninitiated, I'm often skeptical of 2012 presidential polls that rely on partisan sample breakdowns that skew heavily toward Democrats. For instance, the NYT/CBS and NPR polls I broke down last week featured D+6.5 and D+7, respectively. The argument I frequently employ is that 2008's wave election -- a multi-decade high point for Democrats -- produced a D+7 electorate, according to the final exit polls. Barack Obama won that election by 6.5 percentage points. In 2010's midterms, the final exit polls showed the partisan split even at 35 apiece. Republicans won in a romp. Granted, presidential election years typically have significantly higher turnout than midterm years, so it's reasonable to assume that the 2012 electorate will be a hybrid of the last two cycles. Yet many pollsters continue to peg Democrats' advantage in the 2008 neighborhood as they try to predict the Obama/Romney outcome. Jim Geraghty and Jay Cost have each explored this confounding and persistent phenomenon, which consistently benefits the incumbent. As Carol touched on last night, Gallup has released a new national survey of American voters, broken down by party (plus leaners) and levels of electoral enthusiasm. The numbers cast even more doubt on many pollsters' insistence on oversampling Democrats this year. This graph speaks for itself:
For the first time in three presidential cycles, Republicans' intensity outstrips Democrats'. In 2004, Democrats were extremely eager to oust President Bush, with enthusiasm levels approaching 70 percent. They failed -- barely, by a few thousand votes in Ohio. Over the next four years, GOP enthusiasm tanked, hitting a nadir in 2008. Republicans got trounced. The two parties' motivation numbers began to converge as the 2010 election approached, drawing exactly even by election day. Due to ascendant enthusiasm and a strong showing among independents, Republicans delivered a thumping to their rivals. As we're closing in on the final quarter of the 2012 presidential contest, Republican enthusiasm has recovered to its highest point since the successful 2004 election. More importantly, their counterparts' excitement is fading into 2008 GOP territory. Why does this matter? Motivation and enthusiasm are major factors in dictating turnout. Any pollster worth his salt seeks to anticipate turnout factors, and incorporate those clues into partisan sample models. Given the numbers cited above and recent history, it looks like precious few pollsters are making the proper adjustments. Indeed, some are doing the opposite; hence my harsh criticism of the latest NBC/WSJ poll, which relies on a truly ludicrous D+12 (+11 with leaners) sample. This. Will. Not. Happen. A senior Romney campaign adviser reached out to Townhall and offered the following analysis of the NBC/WSJ numbers:
This poll shows Democrats with an 11-point partisan advantage. In the 2008 election (the best D Presidential year in ages) the partisan advantage was -7, the 04 election was dead even. The ’12 election will likely be no worse than -3 to -4. Of the last 62 national polls conducted, dating back to the beginning of April, just three have had partisan advantages as wide as 11-points. Further, among those who voted in 08, their reported 08 ballot was 56% Obama -44% McCain, reinforcing how tilted this sample is.
That last data point is especially damning. For what it's worth, the latest tracking polls from Rasmussen and Gallup show Romney leading Obama nationally by three and one, respectively. NYT/CBS has Romney up one, and NPR has it exactly tied in the battlegrounds (where NBC/WSJ gives Obama an eight point lead).