Skip to comments.Rasmussen botched Washington's Murray v. Rossi poll
Posted on 08/20/2010 10:19:58 PM PDT by dangus
SurveyUSA just released a poll showing that Dino Rossi has a seven point lead over Patty Murray. But RasmussenReports just showed Patty Murray with a four point lead. These poll differences are too great, 11 points, to be an issue of sampling error. So which is poll is correct?
I'm really surprised at Rasmussen being this sloppy. Their poll included the option to answer, "Some other candidate." Four percent chose this option. But Washington state does not allow third-party candidates. Only the top two candidates, regardless of party, compete in the general election. The other major candidates, besides Murray and Rossi were both conservative Republicans, Paul Akers and Clint Didier.
Had these responses been scored as Republican votes, Rasmussen would have found a tie. Each dataset then would also be within the margin of error from the combined dataset, which would give 50% for Rossi, and 46% for Murray. It's likely not a coincidence that the actual election results just happened to be 50% for the three top Republicans, and 46% for Murray.
You don’t necessarily know that all of that 4% would go to Rossi, but your point is valid nonetheless given Washington’s system of primary elections about that 4% being out there that “another candidate” should not have been an option in the poll.
It would be nice if that Survey USA poll would be validated by another poll. We can’t just assume the one that has the results we like are always the most accurate polls.
What about the timing of the two polls? Was Rass done pre Obama visit?
It’s amazing that this especially dull person was elected to the senate to begin with.
Besides, it doesn’t take into account the number of people who will just stay home as opposed to voting for Rino Dossi yet again.
Nor does it take into account the ability of Washington democrats to “find” the votes they need to win - which will also be a factor.
I never understood how Murray ever got elected to anything. She makes Maxine Waters look like a mental giant.
Here’s a video of comments by Patty about Bin Laden
OK, troll. Care to justify calling Rossi, “Rino”?
Rasmussen was released before SurveyUSA’s, but they both state that they were conducted entirely on August 18th.
Rsmussen’s poll was done August 18. Can’t remember when BHO was in WA. Rasmussen also used a larger sample, 750 v. 618. The story on Rassmussen shows a MOE, whereas the SurveyUSA doesn’t. So it’s kind of hard to compare these polls as apples to apples.
It will definitely be a close, fun race to watch.
‘Rino’ Rossi? grow up sonny. go back to your young pals at the DUmp.
“These poll differences are too great, 11 points, to be an issue of sampling error.”
An error of any size can theoretically come from sampling error. Because you are measuring a sample rather than the population, your measure[mean, proportion, etc.] is very likely to differ from the actual measure. If you have drawn your sample correctly large deviations from the true population value occur with small probability, but they can still occur.
What you go on to say is that one pollster failed to list the correct responses in your opinion. It then is not surprising to get different results as a result of giving different samples a different set of answers to choose from.
Every now and then Ras goes off the deep end as they throw a bone to the dims.
Not in my opinion... it’s a matter of fact. You have no other choices besides Rossi and Murray. Suggesting there were other options misled the respondants.
Both polls were conducted 8-18. Obama visited 8-17. MOE is strictly a function of sample size, so the MOE should be very similar.
I thought about the choices when responding, technically could a voter write in someone else? If so, then technically, Rasmussen is correct, but I agree it is still a botched set of responses they offerred voters and 4%writing in someone is very unlikely.
Funny, isn’t it, that when people like the results of a poll, they believe it implicitly, but if they don’t, they seek out all sorts of problems with the methodology.
Well when two polls differ so dramatically, I think anyone would check out why. Some will believe which ever they want to believe; some will try to discover why they disagree.
Nice catch Dangus. You should email Rassmussen about it.