Take that Suffolk poll of MA Senate race with 'grain of salt'
What if you released a poll any nobody believed it? That seems to have happened to Suffolk University, whose new survey of the Massachusetts Senate race rather implausibly shows Republican Sen. Scott Brown leading Democrat Elizabeth Warren by a 49-40 spread. It isn't just commentators on the left who think the pollSuffolk's first since Warren's entry into the racesmells like an outlier; Dave Catanese reports that even Republicans (at least, his unnamed and unquoted sources) think the survey is bunk and that the race is a tossup. (Personally, I think "tossup" is generous to Brown, but we'll leave it at that.)
So what happened here? Well, to start with, Suffolk was one of the weaker firms in Nate Silver's 2010 pollster ratings, ranking 48th out of 62 in terms of accuracy. It's also true that some academic polling operations just have less experienceSuffolk, for instance, only appears to have polled about half a dozen times in 2011and may rely on unpaid students to conduct interviews as part of coursework or major requirements.
But Suffolk's survey construction itself is a problem. They don't get to the actual head-to-heads until the 14th question (not counting the demographic questions they for some reason start the poll with), but what's worse is the content of some of the questions they ask first. Immediately preceding the ballot test, they ask open-ended questions about Brown and Warren which ask respondents to offer "the first word or phrase that comes to your mind" when they hear the candidates' names. Okay, maybe not the worst thing in the world, but then they ask "Does Elizabeth Warren have the experience to be a United States Senator?" and "Is Scott Brown a leader in the United States Senate, or a follower?" Questions like these prime respondents in unpredictable ways and can lead to weird results. It's why good pollsters put the head-to-heads as close to the top as possible and save these kinds of questions for the end. Otherwise you can end up with questionable results... which is exactly what appears to be the case here.
http://www.dailykos.com/story/2012/02/17/1065827/-Daily-Kos-Elections-Live-Digest-2-17?via=user