< .comment-link {margin-left:.6em;}

Massachusetts Liberal

Observations on politics, the media and life in Massachusetts and beyond from the left side of the road.

Sunday, January 10, 2010

Lies, damn lies -- and polls

Either the Massachusetts Senate election is the most volatile in history or there's some serious shenanigans going on among the people who ask the questions and crunch the numbers.

Depending on who's doing the asking, Martha Coakley is 15 points ahead or one point behind Scott Brown in the Jan. 19 election. And she is clearly trailing in the race for Facebook Friends and Twitter Followers, long an accurate bellwether of election results. (I follow Brown, enough said?)

I'm not a pollster or a statistician, but I'm inclined to go with the Globe's findings. Why? They include a link to the full poll, offering everyone an opportunity to see the questions and instructions to the humans who administered the survey.

I've never heard of Public Policy Polling, despite the Wall Street Journal plug on their blog, and then became immediately suspect when I saw the survey is automated, relying on people to punch buttons rather than talk to humans. Call me funny that way.

Either way, there is something increasingly problematic about telephone surveys. We often hear about cell phone bias, excluding the increasingly larger segment of the population that doesn't bother with a landline.

We hear less about Caller ID bias. I know I never pick up a call whether either the identity of the caller or the number is blocked. If you want to talk to me identify yourself.

That's not to say there isn't some common threads in the 16-point spread between the two polls. The Globe-UNH survey notes Coakley and Brown are running dead even among undecided voters who said they were “extremely interested’’ in the race.

But even there, the Globe survey suggests Coakley picks up a few points when you factor in the leaners to each candidate.

About the only thing clear is that Brown has motivated a segment of the population by running aggressively while Coakley has been in a protective crouch. I doubt anyone -- Democrat, Republican or Independent -- will give Coakley's team high grades for running an insprired campaign.

That reality -- whether triggered by incompetence or hubris -- has given Brown the opening to create the impression of momentum. And it remains a fact that perception often trumps reality.

What Coakley needs to do is come out of hiding and engage Brown forcefully on issues that really matter in this Blue Heaven -- health care, Brown's antediluvian positions on waterboarding and gay marriage and his use of scare tactics on terrorism.

Oh, and did I mention the opportunity to be the vote to make or break Ted Kennedy's dream about health care reform?

There are 10 days to go. The long-range forecast suggests the weather will actually be decent (even though I trust long-range forecasts even less than political polls). While Coakley has emerged from the cave, she needs to energize her base the way Brown has energized his.

That's the ultimate bottom line. Given this state's proven political profile, Coakley cannot lose if she manages to generate the excitement and the turnout to match the effort by Brown.

Whether she can and will is the only question that counts over the next 10 days.

Labels: , , ,


Anonymous Charley on the MTA said...

Robo-polling is not necessarily suspect. SurveyUSA is considered one of the most accurate pollsters, and they're all robo-polls, IIRC.

January 10, 2010 9:06 AM  
Blogger Steve said...

I agree with your assessment of the race, and I want Coakley to win as much as you do, but I don't think you're being fair to PPP, a prolific, transparent (read their blog), and also Dem-run outfit: they called the NJ Governor's race precisely, while live-interviewer polls blew it, and there's good reason to think that the Tea Party types who constitute Brown's base are more likely to hang up when they hear "University of" (or, worse yet, "Boston Globe") from a live-interview pollster. Well-conducted robo-polls-- PPP's, SUSA's, and Rasmussen's when they are polling electoral horse-races (they're useless and biased on issues)-- have looked better than live-interviewer pollsters over the last three cycles. So don't shoot the messenger-- especially when the messenger (PPP) wants Coakley to win.

January 11, 2010 9:40 PM  
Blogger Outraged Liberal said...

Charley and Steve, thanks for the comments. As I said, I'm not statistician and I appreciate the feedback.

Personally, I would hang up on a robo call -- assuming I would even pick up in the first place if the number is blocked.

January 12, 2010 5:05 AM  

Post a Comment

Links to this post:

Create a Link

<< Home