Hey, there! Log in / Register
Globe poll: Menino still has comfortable lead
Support Universal Hub
Help keep Universal Hub going. If you like what we're up to and want to help out, please consider a (completely non-deductible) contribution.
Help keep Universal Hub going. If you like what we're up to and want to help out, please consider a (completely non-deductible) contribution.
Comments
Questionable Methodology
David Bernstein brings up the fact that the Globe poll and Flaherty's internal poll are completely at odds. It might have something to do with the Globe poll relying on self-selection for identifying likely and registered voters -- not the norm for political polls.
Link to Bernstein Post
read on
to see two campaign workers debate the poll
Globe poll 20% lead
Flaherty internal poll 10% lead
Wait
Residents? Not voters? Why would they do a poll like that? That has to increase their margin of error right?
Huh?
They did poll registered voters as noted at top of poll results.
Well then why does the article say this?
"conducted the telephone poll of 553 randomly selected Boston residents between Oct. 10 and Oct. 15. Of that number, 438 said they were likely voters. The poll has a margin of error of 4.4 percentage points."
"Likely voters"? That doesn't mean registered. And what about the 115 others who aren't "likely voters"?
It means....
...they are "likely" to actually go to the polls on Nov 3rd. They are all registered voters. Voter apathy is also a quantifiable number in a poll. Sometimes it speaks more loudly then anyhting else.
It means....
...they are "likely" to actually go to the polls on Nov 3rd. They are all registered voters. Voter apathy is also a quantifiable number in a poll. Sometimes it speaks more loudly then anyhting else.
Actually
If you look at the technical results, you'll see that it is just randomly selected Boston adults, who can identify themselves as likely voters..."Are you planning on voting in the upcoming municipal election??" "Sure! Sounds good."
The way the poll is conducted, the polling entity has no idea who they are talking to or if that person is registered since they did not even use a voter file.
Technical Report
Field Period: October 10 to October 15, 2009
Hours: 10 a.m. to 9 p.m.
Median Completion Time: 12 minutes
Sample Size: 553 randomly selected Boston adults (RDD)
438 likely 2009 Boston Mayoral election voters
Sampling Error: +/- 4.4%
Response Rate (AAPOR #4): 18.9%
Bogus Poll
The Boston Globe poll did not target Boston registered voters as is the standard in political polling. For all we know, these people could have not a clue about the municipal election and expressed support of Menino because they see his name plastered all over city-owned property.
I thought the Globe was interested in fairly covering this mayoral election, but if they are not even willing to conduct a poll with standards and integrity, I question their impartiality.
Spin much?
Poll starts by saying it polled 553 REGISTERED voters. Now carry on.
You are wrong
If you actually read the poll, which I did, on PAGE 1 where they provide the information on where they polled, it says they polled:
"553 randomly selected Boston adults (RDD)
438 likely boston Mayoral Election voters"
So pretty much everyone who is over the age of 18 in Boston and has a phone number, was a possible call. It was an added bonus if they are regestered and likely to vote.
Correct
Sample Size: 553 randomly selected Boston adults (RDD)
438 likely 2009 Boston Mayoral election voters
I believe RDD stands for Random Digit Dial. Which means they had no idea who they were talking to. Did not even use a voter list. How can you do a poll like this without making sure you know exactly who you are talking to?
The Globe seems more than a little confused.
The article itself repeatedly says that the globe polled "residents" and even quantifies that around 20% of those residents identified themselves as unlikely to vote. (Of course that didn't stop the globe from including them in the data they reported, but that's beside the point.) The polling graphs may say "registered voters" but that's completely at odds with the information in the article.
Either way, the results are obviously skewed. 80% of residents were likely voters? Even if that was 80% of registered voters, we all know that's ridiculous. The last Presidential election (a pretty rousing contest, I think we would all agree) brought out 61%.
Once again, spot-on reporting and gleaming journalistic integrity from the globe. When are they going to realize they don't have to keep greasing Menino's machine? Change is on the way.
This is why the Globe needs an ombudsman
like it used to have until a few years ago. Oddly, the NYTimes never had an ombudsman before it bought the Globe, but now it has a 'public editor' which is the same thing by a different name. Could we please have ours back?
Good Polls
Good pollsters create a polling model that looks like the turnout for a typical election. This poll relies on respondents to identify themselves as likely to vote. That's where the error lies. Saying that you are going to vote is the "correct" answer, which is why 80% of registered voters (again, a self-selected group) identified themselves as likely voters. That's not even close to reality.
The most depressing statistic for the City
"Only 6 percent of those polled rated the schools as excellent, while 50 percent of residents with children said they had considered moving out of Boston because of the schools, up from 39 percent in May."
Here's who could have been polled...
...by the random-digit dialing that the Globe used:
My twelve year old neighbor...
My non-citizen friend from Norway...
My friend who lives in San Francisco, but still has their 'Boston" telephone number.
According to the Globe poll, we are looking at turnout that will be around 130% on Novemner 3rd. This poll is a complete joke.
The question: Is the Globe incompetent or are they just trying to skewer the results to their preferred candidate, Tom Menino?
The real question
I think the real question is how reliable Globe pasts have been lately i.e. 08 Presidential Primary, 06 Governor's Race, etc. And once you find how reliable they were then maybe also look at if they used the same approach. Let's not take a bunch of pro Flaherty posters opinions and compleely disregard what may turn out to be a good poll.
Totally different types of polling
You can't compare polls like this. First off the Boston Globe poll said that the Mayoral election would have higher turn out than the presidential election. Not to mention that only Boston registered voters will be able to choose the next mayor. Both the presidential and Gubernatoral races included the entire state.
It's easy to find "likely voters"
It takes little or no work to get together a list of "likely voters". The city's elections department has a list of who has voted going back at least ten races. If anyone wants to see who is likely to vote, all he/she has to do is get this list and look at who has voted during the past several elections.
We did this in my race - the benefit is you can target those voters without having to waste time with mailings to people who registered but never voted or those who only voted once or twice during the past ten elections or only in presidential elections.
You can also see the people's political affiliations.
I'd say about half the people who have voted also have their phone numbers attached which helps.
Likely voters
But the Boston Globe did not poll likely voters, or say they got their data from the Elections Dept. They polled randomly selected Boston Residents who, whether they are registered or not claimed they are likely to vote.