Media bias often shows itself in which organizations journalists choose to cite or ignore. A very prevalent form of this bias is selective reporting on polling data--polls that show results friendly to the liberal position like are touted while those that show the opposite are buried.
MSNBC's Chuck Todd, pictured right, is the latest reporter to demonstrate such a bias. He took Rasmussen Reports to task on Twitter yesterday, claiming it is "has a horrible track record and us [sic] proven to be unreliable" and is really "[n]ot a serious polling firm." Todd said he would only report on "numbers from a more reliable pollster."
Apparently one such pollster, in the mind of Todd's cable network at least, is Research 2000. But R2K was recently rated one of the least reliable major polling firms in existence by liberal statistician Nate Silver. R2K was not even accurate enough for the Daily Kos, which officially dropped the firm on Wednesday.
Rasmussen, in contrast, was rated relatively highly in Silver's study, at 15th out of the 63 firms that have conducted 10 or more polls. R2K came in at a paltry 59th.
But R2K's findings have nonetheless been touted on MSNBC in discussing the Nevada Senate race, the very topic on which Todd refused to even consider Rasmussen's findings.
Of course Todd does not speak for all of MSNBC, but where were his protestations when a colleague used poll data significantly less reliable than the unserious, unreliable Rasmussen, with its "horrible track record"?
And if Silver's numbers are not good enough for Todd, consider Rasmussen's actual performance in the last three election cycles. Rasmussen's track record is far from "horrible," as Todd claims. As Greg Pollowitz notes,
he’s consistently one of the more accurate pollsters out there. Here are his results from 2008, 2006 (Senate and governors) and 2004.
Consider the races Nevada: in 2008 Rasmussen’s final poll had Obama over McCain, 55–43. The vote went for Obama, 50–46. (Rasmussen can hardly be accused of skewing Republican there). In the 2006 Nevada governor’s race, the final poll had Gibbons over Titus, 48–44. The election result was Gibbons, 48–46. And in the 2004 presidential race, the final poll had Bush over Kerry, 50–48. The vote tally was Bush, 49–47.
One of the digs lately against Rasmussen is that his 2008 polls are showing a Republican house effect that wasn’t there in other years. Nate Silver has what I think is a fair look at Rasmussen and this development. I assume this is what Chuck Todd is referring to when he calls Rasmussen “unreliable.” The bottom line is we won’t know the answer until November, but if Rasmussen’s past performance is any indication, Harry Reid is in deep trouble. And I think Reid knows it.
