The Globe poll, redux: Pollster responds
Earlier this week, City Councillor Michael Flaherty's campaign lit into a Boston Globe/University of New Hampshire poll that showed Mayor Thomas Menino with a 20-point lead and that while his favorability rating had eroded, it was at 66 percent.
The Flaherty campaign said their own internal poll showed a 7 to 10 point horserace and questioned the methodology of the Globe poll. So we decided to ask the Globe's pollster, Andrew Smith of UNH's Survey Center, to respond. Here's what he had to say in an email to the Reporter:
Regarding the contretemps, this is not unusual when political polls are conducted and is a standard response. It is important to note that the Globe/UNH poll was not intended to be predictive of the outcome of the election, but to give a state of the race in mid-October, and to better understand what the public, and what voters, are considering when making their voting decisions. As the poll indicates, a high percentage of potential voters are unsure about who they might support, and I suspect, if they will even vote, conditions that will become more certain as the election nears and voters finally focus on the choices before them.
Regarding specific methodological issues, random digit dialing surveys that begin with the entire adult population, and then screen down to likely voters, is the accepted standard. Beginning with lists of registered voters, although less expensive, systematically excludes many potential voters (typically 40% - 50% of all registered voters) whose phone numbers are not available to list compilers. This poll was also designed to be comparable to an earlier poll conducted in May this year which asked several of the same questions.
I do agree that a percentage of those who said they would likely vote in the election will end up staying at home on election day. This is a common phenomenon well known to survey researchers, which many have tried to crack, but none successfully. People over-report their intentions of voting because voting is seen as a good thing in our society, and who [wants] to admit being a bad citizen? There is some evidence that non-voters systematically exclude themselves from surveys by refusing to participate at the start.
The important factors in this poll, to me, are the very high job approval and favorability ratings that Menino has, and the high percentage who feel Boston is headed in the right direction. Elections are typically referenda on the incumbent, and these factors make Menino a very formidable force in the upcoming election.