Rasmussen vs. Quinnipiac – Blumenthal vs. McMahon

Hold on here a second. Will someone in the media please take a look at the polling methodology between the Quinnipiac and Rasmussen polls concerning the US Senate race in Connecticut? Why are the results so different?

The Rassmussen poll was taken on May 18 and asked questions of 500 likely voters. The Quinnipiac poll was taken on May 24 and 25 and asked questions of 1159 registered voters.

So we’re talking about a six to seven day gap, and the results are that different? The Rassmussen poll was a statistical dead heat with Blumenthal leading a direct race with McMahon 48 to 45 percent.

The Q poll released today has Blumenthal leading the GOP-selected McMahon 56 to 31 percent.

I don’t have the time to get into all of the details, and I don’t have the crosstabs for the Rassmussen poll, but instead of just touting the hey-look-McMahon-is-in-a-freefall story, can someone in the media explain the difference and why/how this happened?

I’m pretty certain nobody in Connecticut media will step up to take a look, so I’m calling Ed Morrissey from Hot Air and will ask him to dig in.

Update: Morrissey at Hot Air takes a look.

7 replies
  1. chris-os
    chris-os says:

    1. Polling more people =higher accuracy.

    2. The Blumenthal wave crested and has went out to sea.

    3. The R's eliminated their only viable candidate-give CT voters credit, they (from both parties) cannot fathom sending someone to the US senate that isn't qualified to be 1st selectman of a small town.

    • Steve McGough
      Steve McGough says:

      If I remember my statistics class, the total poll sample should not matter after you hit a certain number of people, so I don't think that explains it. Your confidence rate would get better with more people asked, but I would think 500 would be a good sample size if they were likely voters.

      I'm more interested in some research on the details for each of the samples. What was the weighting like for both polls – Democrats/Republicans/Other – as compared to the actual percentages within the state? Also, the locations/home towns would be interesting to know.

      At this point, a good portion of the poll results would have been driven by media coverage, and although I was out of the loop completely Friday through Sunday, I can not figure out how the spread went from 3 to 25 points in a matter of seven days.


  2. Lynn
    Lynn says:

    Steve,thanks for bringing this question up.  There is something very wrong to make that drastic a change.  I agree there must be totally different methodology used in the two polls. Unfortunately a poll of " likely voters " doesn't cut it concerning election results. Q-poll sounds a little more accurate to me.  Since print media no longer does any digging on its own, Linda gave NY Times their story. Now she will take the heat for being the messenger.

  3. chris-os
    chris-os says:

    Poor Linda, has to take the heat!

    She hired Holloway, a protege of Rove, a specialist in attack politics.

    They selectively clip a part of a speech-the remaider of that speech has him saying that he served during the time of…

    Think CT voters are stupid? They know her background-and I keep saying, if she did not have millions, she would be laughed out of running for a counsel seat in a small town!

  4. Lynn
    Lynn says:

    Chris, I am not sorry Linda is taking the heat. I am well aware of her campaign strategists. You get what you pay for. She bought her nomination. I am still nursing a broken heart. I have been a Rob Simmons supporter from way back. Three of us worked at his campaign headquarters the Wed. before the Convention.  His campaign staff were all in their 30's. I have never been involved in a more enthusiastic campaign before. Back room switching of delegate support at the Convention was easily observed. Too bad. Are you from Old Saybrook?

Comments are closed.