Now that the confetti has been swept up, and tears (of joy or anguish) have largely been dabbed dry, time for a little dispassionate analysis. As the resident poll guy here, I have been fairly steadfast in my criticism of Rasmussen Reports, as have several of my colleagues.
Given that Daily Kos partners with among the most accurate pollsters in the game (Public Policy Polling), it has always irked me to no end that the political media has always referenced any PPP poll as a "Democratic poll" (despite actually having a slight GOP house effect during this cycle). Meanwhile, Rasmussen (whose rooting interest, at this point, seems beyond discussion) is never categorized similarly. Indeed, one of the "serious journalists" actually described Rasmussen as "a pollster that liberals claim leans to the right."
Well, now we have results. And here is the evidence:
- RASMUSSEN IS BIASED: Rasmussen polled over 75 races from the middle of October until Election Day. Let's narrow that down to their most recent polls in each race (since, after all, they polled some races multiple times). So, looking at unique races, we get down to 57 contests. The House of Ras overestimated the margins in the race to the advantage of the GOP in 46 out of those 57 contests (81%). In an ideal world, of course, a pollster would get their numbers wrong pretty evenly. For example, with our pollster (PPP), their numbers wound up overstating the Democratic performance 17 out of 32 times during the same time period, and the GOP 14 out of 32 times (they hit IL-Sen right down to the tenth of a point). It cannot be attributed to mere coincidence that Rasmussen's polling overstated Republican performance over three-fourths of the time.
- RASMUSSEN IS NOT TERRIBLY ACCURATE: While PPP's polling had a few misses (as all pollsters will), they were largely confined to the impossible-to-predict Alaska Senate race and a couple of House races. In a tribute to their accuracy, 17 of the 32 polls PPP has conducted since mid-October led to them coming within 3% of the final margin (53%). How did Rasmussen do on that score? Twenty-one out of 57 polls (36.8%) met that same standard. They also get the reward for dropping the biggest bomb of a poll this cycle. On October 20th, they reported to a shocked America that the GOP wave had even reached the impenetrable longtime Hawaii Senator Daniel Inouye. The House of Ras had the Hawaii Senate race as a 13-point race, with the man that has represented Hawaii since statehood up only 53-40. On Election Day, of course, Inouye won by...53 points (75-22).
- RASMUSSEN IS NOT TERRIBLY BOLD: Maybe the biggest curiosity for me in this election cycle is how timid they were at the close. Only three polls in the last day. Only 11 in the last three days. To place that number in context, place it in comparison to PPP. PPP hit fewer polls over the final three weeks (57 to 32), but hit twice as many polls as the House of Ras in the final two days. Unlike Rasmussen, who reserved the right to hide behind the "late breaking dynamics" excuse if they chose, PPP put its reputation on the line by calling their shot at the last possible second. Remember, of course, that Rasmussen did the same thing during the primaries. Most notably, they also did it during the Massachusetts Senate race, when they refused to poll the race in the last week, but inexplicably ran instead a national poll about who the country wanted to see win the race.
Nate Silver of 538/NYT fame crunched his own set of numbers on the House of Ras last night. He included the work that Ras subsidiary Pulse Opinion Research did for Fox News, and he found similar results:
Of the roughly 100 polls released by Rasmussen or its subsidiary Pulse Opinion Research in the final 21 days of the campaign, roughly 70 to 75 percent overestimated the performance of Republican candidates, and on average they were biased against Democrats by 3 to 4 points.
Every pollster is entitled to a bad cycle now and again -- and Rasmussen has had some good cycles in the past. But their polling took a major downturn this year.
Clearly, one of the consequences of last night's midterms is that a number of changes are going to come to America. In the insular world of the political press, it is high time for one long overdue change: the ability to recognize Rasmussen for what it is. And this cycle proved is that Rasmussen is a partisan outfit that does not necessarily produce great results, and puts its thumb AND forefinger on the scale for its favored party.