Not surprisingly, the new WPRI poll is being touted as the possible end of Patrick Kennedy’s political career.  I’m not sure that’s entirely true because the survey design doesn’t actually give a very accurate measure of public sentiment.

The poll shows 31 percent of those interviewed said they would “consider another” candidate and 28 percent said they would “vote to replace” Kennedy. Those who would re-elect the eight-term Congressman came in at 35 percent. Five percent weren’t sure.

What does “consider another” candidate actually mean?  I suppose I would also consider someone else, but when it comes down to voting for either Kennedy or Loughlin, I’m going to vote for Kennedy. That’s my frustration with the poll.

Whenever there is an illusory alternative, people find it more attractive than if given a hard choice.  When questions are asked about whether or not people would vote for a Republican of Democrat in the next election, or “consider another,” the results are largely meaningless.  

The polling firm Fleming & Associates (is that the same Joe Fleming who is on Paiva-Weed’s payroll?) should have asked “if the election were held today, for whom would you vote: Democrat Patrick Kennedy, or Republican John Loughlin?"  Maybe then we would have a better understanding of how people felt about Kennedy.

Additionally, there is an over sampling of Republicans, and under sampling of Democrats and Independents.

The poll was conducted by phone January 27 through January 31, 2010. Reflecting the political canvas of the state; 39 percent of those polled considered themselves Democrat, 17 percent Republican and 41 percent Independent. Statewide statistics -- 500 registered voters -- comes with a 4.4 percent margin of error.

In a September 2008 Brown University survey, Patrick Kennedy’s job performance was as follows: 

15. How would you rate the job Patrick Kennedy is doing as U.S. representative? Excellent 11.1%; Good 35.6%; Only fair 26.6%; Poor 20.2%; Don’t know/No answer 6.5%

While only 47% of voters said they thought Kennedy was doing an excellent of good job, he got almost 69% of the vote in the November 2008 general election. 

What all this means is that poor survey design, or surveys taken out of context, means poor survey results.

I suppose that if you wanted to create a poll that reflects what the media frame currently is, that a specific Democrat named Kennedy may lose he seat as did Coakley lost another Kennedy seat in Massachusetts, this is the perfect poll.  If you want an accurate result of what people in the state (or more specifically, likely voters in District 1) are actually feeling about Kennedy, this poll fails to do that in any meaningful way.