For more than four years now, I’ve been arguing here that polls from Rasmussen Reports should be taken with a huge grain of salt.
Noted analyst Nate Silver put it best when he said this last year: “My advice would be simply to disregard the Rasmussen Reports poll, and to view their work with extreme skepticism going forward.”
And now Alan Abramowitz, a political science professor at Emory University, is out with THIS INDICTMENT of Rasmussen:
In 2012 as in 2008 and 2010, Rasmussen has been the nation’s most prolific polling organization at the state level. Almost every day Rasmussen produces two or three new polls of key battleground states in the presidential election along with their national tracking poll. But in addition to being the nation’s most prolific pollster, Rasmussen gets a lot of attention, especially from conservative media outlets and pundits, because its polls consistently produce results more favorable to Republican candidates than the overall averages — results that frequently don’t match the actual election results very well.
So how is Rasmussen doing this year? Along with its national tracking poll that has typically showed Mitt Romney with a lead that is two to three points larger than the overall average, Rasmussen’s polls in the battleground states have also had a consistent Republican lean. I compared Rasmussen’s latest results in the nine key battleground states — Ohio, Florida, Virginia, Colorado, New Hampshire, Iowa, Wisconsin, Nevada and North Carolina — with the overall averages from the HuffPost Pollster — an average that includes Rasmussen polls. In every case, Rasmussen’s poll showed a better result for Romney than the overall average.
In 2012 as in 2010, Rasmussen is producing polls in key states that are significantly more favorable to the Republican candidate than an average of all polls in those states. Perhaps this time Rasmussen will prove to be more accurate than the average. But outliers like Rasmussen usually turn out to be wrong and Rasmussen’s track record certainly shouldn’t give election watchers much confidence in their results. Maybe that’s why Fox News stopped using Rasmussen for their state and national polling this year. Producing polling results that please one side or the other may work for a while, but eventually journalists and voters catch on to the fact that those results don’t seem to match the numbers on election night.
Meanwhile, Sam Wang of the Princeton Election Consortium SAYS this morning that the likelihood of President Obama getting re-elected now ranges from 86 percent to 95 percent — which is down a smidgen from yesterday.
Wang’s current projection of electoral votes has Obama with 295 and Mitt Romney with 243, which is almost exactly the same as the aforementioned Nate Silver sees it.