According to Rasmussen, the race for Ohio Governor went from 46% Kasich to 45% Strickland (essentially tied for the past two months) to suddenly a five-point race in Kasich’s favor. However, the poll has a margin of error of 4.5% (which is actually rather high). Given that Kasich’s numbers have shown no movement for months (according to Rasmussen), and Strickland’s movement is within the margin of error of his results in the past two Rasmussen polls and all other polling shows Strickland with the momentum, it’s likely that this result is statistical noise and not real movement.
Regardless, Rasmussen just cemented its reputation as an outlier in Ohio polling and amplified it.
The Carpetblogger pointed out that Nate Silver at FiveThirtyEight.com ranked the Ohio Poll, by the University of Cincinnati, as one of the highest rank polling outfits for reliability. Rasmussen came far behind.
And the Carpetblogger’s own Tweets about the Rasmussen poll shows that is it clearly at a polar opposite of the result found by likely voters by the Ohio Poll
- Keeling Tweet: “Strickland is only getting 73% of the African-American vote. That has to get better, right?”
- Ohio Poll: Strickland has 95% of the African-American vote (see pg. 6)
- Keeling Tweet: “Kasich is winning every age group except the youngest. And they don’t vote.”
- Ohio Poll: “Strickland beats Kasich by 11 points with voters 65 and older” (This is the most reliable age group to vote.) The Ohio Poll does agree with Rasmussen about the youngest voting population, but it shows all other age groups in a virtual tie.
- Keeling Tweet: “Kasich is winning Independents 53-27. That’s huge.”
- Ohio Poll: “Strickland leads independents by 44%-30%”
There are other examples that show a huge disparity between this most recent Rasmussen poll and the recent Ohio poll. Rasmussen claims that Kasich has a double-digit lead with male voters; the Ohio poll shows a dead heat with male voters.
Rasmussen claims that Ohioans approve of Strickland by only 43% with 55% disapproving; the Ohio Poll had essentially the opposite with 55% of Ohioans approving of Strickland’s job performance and only 35% disapproving.
There are fundamental questions about Rasmussen’s methodology that calls into question the reliability of their polling data:
- According to Nate Silver, Rasmussen does not call cell phone users (who tend to trend younger and more progressive). The recent Ohio Poll did poll cell phone users.
- Rasmussen continues to show the smallest party identification for Democrats among all adults of any national polling outfit.
- Rasmussen has continuously shown, by far, the largest pro-GOP generic congressional ballot result of any national polling outfit.
- Rasmussen uses none of the established and accepted polling practices to adjust for errors that arise from raw polling data (calling and polling whomever answers.)
- In Rasmussen’s case, it’s even worse in that since they use an automated dialer to conduct their poll, they have little idea who in the household was actually polled. They rely entirely on the honesty of their anonymous respondents. As you can imagine, this might lead to some questionable results.
- Nate Silver has noted that Rasmussen’s “house effect” is more pronounced than it’s been in the past, meaning it’s been even more of an outlier in polling races when compared to other polling outfits than its been in the past.
- In fact, Silver has noted that this year, Rasmussen has the most pronounced “House effect” (favoring candidates of a particular party consistently in race after race) more so than any other polling outfit in the nation (Democratic or Republican alike).
If I were the Carpetblogger, I wouldn’t get too excited about a poll that could be showing nothing more than statistical noise. Given the facts that we’re seeing on the ground, Rasmussen’s claim of a five-point drop from Strickland defies logic and reason, particularly in the face of the strong jobs report in April and the fact that Kasich’s campaign has been doing nothing but taking it on the chin over Wall Street. At best, maybe the race hasn’t changed much in the third month in a row.
But a pro-Kasich bounce of five points just isn’t credible. I wouldn’t be at all surprised to see a sudden Strickland “bounce” in Rasmussen’s next poll where things return more in line with what other polling outfits are showing.
Regardless, it’s clear that Rasmussen continues to be an outlier poll that shows a pro-Republican bias when compared to other more respected polling organizations in Ohio. You know it’s bad when the only poll that has ever come close to Rasmussen Reports was the one conducted by the Ohio Right to Life.