By now, many of you can probably recite by heart my cautionary tale from the sample presidential poll that takes place almost every week. For the uninitiated, I am often skeptical of 2012 presidential polls, which are based on a partisan split and lean heavily toward the Democrats. For example, the NYT/CBS and NPR polls I reported last week showed D+6.5 and D+7, respectively. An argument I often operate is that according to the final exit polls, the 2008 election wave – the most essential point for Democrats in many decades – created a D+7 electorate. Barack Obama won this election by 6.5 percentage points. In the 2010 midterms, final exit polls showed a partisan split of up to 35 people. Republicans won quickly. It’s true that turnout is typically much higher in presidential election years than in midterm years, so it’s protected to assume that the 2012 electorate will be a hybrid of the last two cycles. However, many pollsters are still assessing the Democrats’ advantage in the 2008 neighborhood when trying to predict the Obama/Romney outcome. Jim Geraghty AND Jay Cost each of them has explored this confusing and persistent phenomenon that continually benefits the incumbent. As Carol mentioned last night, Gallup published fresh nationwide survey American voters broken down by party (plus people on a diet) and level of electoral enthusiasm. The numbers further cast doubt on many pollsters’ insistence on oversampling Democrats this year. This chart speaks for itself:
For the first time in three presidential cycles, Republican intensity exceeds that of Democrats. In 2004, Democrats were very keen to unseat President Bush, with enthusiasm levels as high as 70 percent. They failed… hardlyBy several thousand votes in Ohio. Over the next four years, Republican Party enthusiasm waned, reaching a nadir in 2008. Republicans were defeated. The two parties’ motivational ratings began to converge as the 2010 election approached, and on Election Day they evened out. Thanks to growing enthusiasm and mighty showings among independents, Republicans made a solid showing of their rivals. As the final quarter of the 2012 presidential election approaches, Republican enthusiasm is at its highest level since the successful election of 2004. More importantly, their counterparts’ excitement is waning and moving into 2008 GOP territory. Why does this matter? Motivation and enthusiasm are the main factors in dictation to turn out. Any pollster worth his salt tries to predict turnout factors and incorporate those cues into partisan sampling models. Given the numbers cited above and recent history, it appears that few pollsters make appropriate adjustments. Indeed, some do the opposite; hence my edged criticism of the latest NBC/WSJ poll, which is based on a truly absurd D+12 sample (+11 for slim people). This. Will be. NO. Become. He was contacted by a senior adviser to the Romney campaign Town hall and provided the following analysis of the NBC/WSJ issues:
This poll shows Democrats with an 11-point partisan advantage. In the 2008 election (the best D presidential election in ages), the partisan advantage was -7, and the 04 election they were even dead. The 2012 election will probably be no worse than -3 to -4. Of the last 62 national polls conducted since early April, only three had a partisan advantage as enormous as 11 points. Next, among those who voted in ’08, their reported vote in ’08 was 56% Obama -44% McCain, confirming how lopsided this sample is.
The last data point is particularly damning. For what it’s worth, the latest surveys tracking from Rasmussen AND Gallup show that Romney leads Obama nationally by three and one, respectively. NYT/CBS has Romney’s lead by one, and NPR is exactly tied in the battlegrounds (where NBC/WSJ gives Obama an eight-point lead).
