Rassachusetts: Why a Poll May be Terribly, Horribly Wrong -- And Why Democrats Should be Worried Anyway

OK, here's that Rasmussen poll on the Massachusetts Senate special election. It shows Martha Coakley leading Scott Brown 50-41, a 9-point margin, eerily replicating the expectations of the Democratic strategist who told Dave Weigel: “I’d guess that, being Rasmussen, it’ll have a 10-point race.”

Although a 9-point margin -- or closer -- seems entirely possible to me, Rasmussen has arrived at that number in something of a strange way. It's pretty easy to back out Rasmussen's turnout demographics, and they're showing an electorate which is 21 percent Republican, 52 percent Democrat, and 27 percent other. Although there are lots of different ways to ask about party identification, typically that's not what we see in elections in the Bay State, as the number of independents is usually much higher (43 percent of Massachusetts voters were independent/other in 2008, and 51 percent are registered as independents). They're also showing an electorate that is 39 percent liberal, 34 percent conservative, and 27 percent moderate; that compares to 2008 exit poll demographics of 31 percent liberal, 19 percent conservative, and 49 percent moderate.

So Rasmussen's theory on this election, basically, is that the people in the middle won't bother to show up; there are many fewer independents and many fewer moderates in their sample than you usually get in Massachusetts. Instead, it will be a race between the bases. That could be a good theory, or it could be an artifact of their sample design -- one thing that generally seems true of Rasmussen and some of the IVR pollsters is that they capture a hyper-informed and hyper-partisan electorate. (To wit: Rasmussen shows Coakley getting just 21 percent of the "other" vote -- but 24 percent of the Republican vote.)

By the way, that's not necessarily meant to imply that Rasmussen is lowballing Coakley's number. It could be that they're low on Brown instead. Or that they have two wrongs which more or less make a right. Or that they have the race completely nailed. Or that they've completely flubbed it up.

But, if I had to set an over-under on this race, it would be above 9 points for Coakley, especially given the earlier polling. If Coakley were to hold the 58 percent she got in the Suffolk and WNEC polls, for instance, at Brown got all the undecideds, that would imply a 16-point margin of victory. Or, if you take Rasmussen's 9-point margin and add 4-5 points to it, which has been roughly the magnitude of their house effect thus far this year, that would imply a 13-14 point win.

So something in the broad range of about 15 points seems about right -- as an over-under. But it's also an over-under with an exceptionally high error margin, because special elections are so difficult to predict. A 30-point drubbing by Coakley wouldn't surprise me; nor would a race that kept us up late on Election Night.

And frankly, if I were either party and my internal polling showed a 15-point margin, I'd still be thinking about putting some money into this race. Special elections in many ways resemble presidential primaries, and polls are off in primaries by an average of about 7 points. That would imply something like a 3-5 percent chance of a Brown victory, which feels about right. But considering how consequential that 5 percent could be -- the probable collapse of health care reform -- it's something worth hedging against if you're the Democrats, or taking a flier on if you're the G.O.P.