I would note that the youth vote goes to Reilly with 46% of the youth vote. This is surprising, given my anecdotal evidence. I’ve only met 2 people under the age of 25 that support Reilly. Nevertheless, the number is there, with 435 likely Democratic Primary voters surveyed.
It’s also interesting to note that Deval has pretty steady support in the three broad regions of MA: West/Central, Boston Area and South East MA and Cape taking 37%, 36%, 35% respectively. Reilly and Patrick are essentially tied for the suburban vote with 34% to 33%.
One more thing, Reilly has about 39% of the Independent vote to Patrick’s 30%, but with margin of error that could easily be tied.
All these numbers will change, I’m sure, as Gabrieli’s ads become more effective and frequent and as the campaign season really gets into high gear after labor day. Also, thousands of households were contacted on June 17th statewide canvassing for Deval, which this poll does not take into account.
Gabs and Deval will kick each other and be kicked by others through out the summer. What more is their to say about Reilly? (Ernie’s gonna knock him)
<
p>
By Labor Day either Gabs or Deval (Gabs most likly) will be a distand third behind Reilly and and the other. Reilly will win on TRUST.
<
p>
Even though Ernie believes him to be an opportunistic fraud. Or just plain stupid. Which is better?
We’re getting into the part of the campaign when polls like this are useful, but only in detail. The least useful part of this poll is the “if the election were today, who would you vote for?” question. What we should be looking for in polls, now, is name recognition for each candidate, favorable/unfavorable/neutral numbers for each candidate, and how those shift from one poll to the next. We should also be paying attention to their polling universe – are they polling registered or “likely” voters, are the polling primary or general election voters, etc. With that level of detail, we can find out how people are starting to pay attention, how they’re reacting to the candidates, and how news events or campaign ads or strategies are affecting things – or at least make good guesses about those.
<
p>
But “who would you vote for today?” – still mostly useless.
because then you have some idea of the “news cycle””events” or anything that happened to influence the result.
<
p>
This poll was conducted 6/16-6/18 and my initial reaction was that it is devastating to Gabrieli because news of his $15.36 million spending cap had plenty of time to “sink in” with respondents.
<
p>
Even my FOXNEWS lovin neighbor thought that number was outrageous and even added that he hoped Kerry Healey didn’t spend money like that – in his words “it’s obscene”
Be careful with the sub-trends like “youth vote” and “senior vote” coming out of this poll. They are far less statistically reliable than the combined result. For example, if the margin of error of a poll is listed as 3%, that number is based on the entire sample size and applies only to the final split. Any breakdowns on a specific demographic (i.e. youth vote, senior vote, etc) were done on however many say of that demographic happened to be included in the existing sample, and the margin of error can be much larger. In the case of youth voters, there probably weren’t very many in the poll, because most likely primary voters aren’t youths. So in the absence of further information, I would guess that the “youth support” statistic in particular is pretty unreliable.
…since Patrick had the same 36% in the same poll two months ago. (Although, on further reflection, Patrick may have dipped during his run of bad news in May and come back to roughly the same level thanks to the convention win story.)
<
p>
And though Cos is right about the horserace number being less useful than the screens the polster used or some of the internals, it is ultimately a horserace number — the Primary Day result — that determines who wins and goes on to the general. And on this front, the poll is bad news for Tom Reilly.
<
p>
In his eighth year as a statewide elected official, the AG is stuck at 31%. Why haven’t more of his voters found him by now? He’s had a long time to make an impression on them.
<
p>
What are Reilly’s advantages over the next three months? Exciting ideas? Personal magnetism? Campaign cash? Field organization? No, no, no, and no.
<
p>
The electability argument? That hound ain’t huntin’: the most recent head-to-head poll had all three Dems beating Healey by 8-10 points.
<
p>
The support of elected officials and unions? We’ll see. But IMHO there is neither the depth of warmth toward TR or a compelling “insider” self-interest in electing him, nor a fear of or antipathy toward Patrick or Gabrieli that would truly motivate legislators, mayors, or unions to pull out all the stops for Reilly. (And the other candidates have significant support from politicians and unions, too.)
<
p>
OK, he’s the “I’ll cut your taxes” Democrat. This MAY be the more popular place to be in November. But among Democratic primary voters?
<
p>
I don’t believe our primary will be a referendum on tax cutting. It might not help Reilly even if it were.
<
p>
It says here that this is a Patrick-Gabrieli race by Labor Day.
Does anyone have a sense of the different methodologies used by these polls? For whatever reason, the Survey USA polls (which this is) have consistently seemed to be more favorable to Patrick and more unfavorable to Reilly than the other polls. (A good summary of the polls is here: http://en.wikipedia.org/wiki/Massachusetts_gubernatorial_election%2C_2006). Do you know why this is? How does each define the universes polled?
<
p>
My take is that if you compare this poll with the last Survey USA poll (May 1-3), it suggests no or little convention bounce for Patrick, a very slight drop for Reilly (down 2), a slight gain for Gabrieli (up 4), and a small drop in undecideds (down 2). Given the margin of error, it’s hard to make too much of these minor variations in polls at this point. But it does suggest somewhat bad news for Patrick, given how much he was counting on momentum from the convention. Otherwise, pretty much status quo.
<
p>
(Note: I’m a Reilly supporter).
http://www.surveyusa.com/methodology.html
A friend pointed this out to me in the methodology:
It’s a recorded call!
<
p>
What does that mean? Lots more people will hang up on this poll than on human calls. And the refusals aren’t random: People who actively support a candidate are much more likely to want to answer a poll. People are are paying a lot of attention to politics are much more likely to want to answer a poll.
<
p>
That explains why there are so few undecided in this poll. Most polls understate undecideds, by pressuring people (however subtly) to pick someone. But a recorded call would do so even more.
<
p>
That explains why Gabrieli is doing poorly: Most of his “supporters” at this point are people who don’t know anything about the race, aren’t paying much attention, but have seen his TV ads. They’re much less likely to answer a poll than people who are actually politically active and committed to a candidate.
These results are of “likely primary voters” already, so whether or not it’s an automated poll does not enter into the analysis. That explains why there are few undecideds — people who have already decided are more likely to vote.
<
p>
Automated polls are often much better than human-voice poll because they can get a larger sample size more cheaply than can other polls.
If there is bias introduced because of systematic non-response, then the increased sample size only means that you have a smaller margin of error for your incorrect poll. Cos is right about automated polls and bias.
<
p>
But (living up to my name here) anyone who cares that much should be able to test this out by looking at survey USA polls compared to other polls for important 2004 races. Were they on target? Are there contested Dem primaries that someone could look at?
<
p>
Also, is there research into these automated polls (not from the companies that are selling them)?
<
p>
But to Sco’s point, “better” should speak to a poll being accurate. A larger sample size speaks to a poll being precise. (An example is if someone says it’s really hot out in the 80’s or low 90’s and I say it is exactly 147.45 degrees out, I am more PRECISE but the former person is still more accurate – and clearly his analysis is “better.”
Comments!
Automated Pollsters killed the competition in 2004.
<
p>
Here is what SUSA has to say for itself.
<
p>
I can’t imagine the response rates are much worse than with human polls. If people want to be polled, they’ll stay on the line if it’s a human or a recording of a media-type. It’s true that automated polls may reduce the number of undecideds (“push 3 for undecided” comes at the end of the list of candidates), but there will be no undecideds on election day, so automated polls end up coming closer to actual election results.
<
p>
That makes no sense. First of all, there will be some people who don’t vote, or don’t vote for that office, because they’re undecided, so factually that’s incorrect. Practically, you’re right, most of the people who are undecided now will make up their mind by or on election day. But what makes you think they’ll break the same way as people who have already made up their minds? People who are decided and undecided are clearly different in some way, and that may mean they make different choices about the candidates. Pretending that leaving out the undecideds in a poll now means we’re closer to the actual election, is just getting a false sense of confidence about something we should have a lot more doubt about.
At the voting booth, no one will be voting for “undecided”. Final election results will not contain a percentage for “undecided”.
<
p>
Of course there will be people who are actually undecided and don’t show up, or there will be people making their decision at their polling place.
You have far too much faith in “likely voters”. Just because someone is “likely” to vote, doesn’t mean they’re especially likely to be paying a lot of attention to the governor’s race three months in advance. When we say most voters aren’t paying attention yet, we do in fact mean “most voters who are going to vote in the primary”. The fact that this poll was of likely voters has absolutely no bearing on what I said.
I don’t have faith in likely voter calculations. In fact, I would argue that the way you determine who is a likely voter may invalidate the results of a poll if done poorly.
<
p>
That said, a poll of likely voters rather than registered voters is more likely to contain people who are paying attention to elections. I would argue that it’s winnowing the survey down to likely voters that is responsible for the effect you describe, and not the fact that SUSA uses automated polls.
<
p>
Also, when a person asks you if you’re going to vote, you’re more likely to respond with what you think the person wants to hear — yes. People are more likely to be honest if they don’t think anyone is listening. Advantage auto-poll.
Or is it just your guess? Just because you think that’s probably the case – that people will act that way – doesn’t mean it’s what will happen. That’s why we should look for evidence.
<
p>
Frankly, people are pretty honest when answering surveys to human callers, and I doubt there’s much of a difference with auto calls. The issue is about bias from low response rates.
<
p>
Do you really think only 9% of people who will be voting in the primary are undecided?
We can go back and forth on this all day. The fact of the matter is that people do lie to pollsters. Look at the results of any gay marraige amendment poll versus the final vote in states where it was on the ballot. It happened here in 1990 when the polls showed that Bellotti was running away with the Dem Primary even days before the vote. Pollsters could only figure that people were too embarassed to admit that they were voting for Silber.
<
p>
Do you have any evidence that automated pollsters have a lower response rate than human-voiced calls? Or is that your guess?
<
p>
Also, look at the phrasing of the poll. The question asks the respondent to imagine themselves in the voting booth. There is no “undecided” in the voting booth, so they may feel forced to pick a candidate. Does that mean that this will be the candidate they will vote for in September? No, but it’s a decent measure of who people see as their current favorite.
The Patrick 8-point increase was the biggest change since the last State House News Poll (21 since the last poll listed there), while Reilly was down 1 and Gabrieli down 6. It implies that Reilly didn’t gain or lose much from the whole process, while Gabrieli’s ad-induced jump may have been undone a bit by his convention antics and Patrick’s win. Now comparing it to begining of April, pre-GabAd Blitz, it is very static: is that the comparison you were making?
When the same pollster asked the same questions two months ago, the result was statistically the same as this recent one. But these polls are two snapshots two months apart, and yes, Gabs’ ad buy and Patrick’s run of bad news in the intervening time no doubt caused some fluctuation in between.
<
p>
So was there a bounce compared to one month ago? Yeah — you just have to look at other pollsters’ results from that time period. But compared to two months ago, the horserace numbers are surprisingly stable.
<
p>
Again, Cos is right — we should look at favorable – negative – never heard of numbers over time among those most likely to vote in the primary.
I agree with most of what you wrote, but I was pointing out that the April poll was the only one to that point to show Patrick in the lead, and that looking for a “convention bounce” it’s more reflective to look at the May one. If the April one had been more like the general trend it would have been a better marker to compare to.
<
p>
Though I have no idea why no public polls were released for over a month that included most of Gabrieli campaign and the nominating convention. The only one I saw in that time was internal Gabrieli one.
<
p>
Also, I do wish this one had approval ratings, as those are more useful, though I think we’re getting to the point that horserace are starting to be useful (below 10% undecided is pretty good for 3 months before i think)
You can’t compare this poll with the May Statehouse News poll because the methodologies are different. You can compare trends, but it’s just inaccurate to say that anyone went up or down from a poll done by another outfit. In this case it doesn’t make sense because the Statehouse News poll doesn’t winnow its survey down to likely voters, whereas Survey USA does.
I was comparing the April May and June SurveyUSA polls, all three of which were just over 400 likely democratic primary voters.
<
p>
Patrick – 36% 28% 36%
Reilly – 33% 32% 31%
Gabrieli – 19% 29% 23%
Undecided – 11% 10% 9%
for comparison:
<
p>
Patrick 37
Reilly 47
Undecided 17
How the heck did I miss that one?
That argument doesn’t make a whole lot of sense to me. Granted, we all want to read things into the results in favor of our candidate, but we gotta draw the line somewhere.
<
p>
<
p>
Because they’re sampling candidates. Gabs and Patrick are still unknown, and your average voter is unaware of Patrick’s history or Gabrieli’s platform. Just like with any splashy new candidate, people are taken with the charisma and project their desires on the new guy(s). These numbers haven’t settled yet.
<
p>
Reilly is a safe candidate, in both good and bad senses. He doesn’t have that new-candidate sheen, but nor are there questions about “what don’t we know”. His impression is a reliable, honest public servant, which doesn’t stand up against a dazzling up-with-people theme. But I think most voters are about more than “impressions”.
<
p>
<
p>
First of all, he has more cash than Deval, $4 million. He doesn’t have far-fetched “exciting” ideas, only practical ones. He has a field organization, it’s just being activated. Why do you think he got 27% of the delegates without bothering to work the caucuses?
<
p>
<
p>
They always say that at this point, yet we always lose. I’m sure Deval’s people would like to cliché that argument out of existence (and it’s bad enough we feel a need to import Southern aphorisms into our national campaigns to sound down-home, but can we at least keep our state campaigns decent?) but it’s very real. Especially when facing a millionaire.
<
p>
This is a great attempt to play the expectations game — because he’s not where I say he should be, he’s losing! The novelty factor does not make someone a candidate, and it’s already wearing off for Deval. The hip new thing is Gabrieli for the moment, and that’ll wear off. I look forward to comparing these candidates on their merits, not their rhetoric.
How about Tom Menino, Bob Travaglini, Sal DiMasi, Guy Glodis, Marty Meehan, et al.? It had something to do with field, but it was all about depending on established pols who could twist arms (particularly with payroll patriots). Give the field credit for getting what they could out of those resources, but winning in a primary is a different world – just ask Tom Birmingham who had similar connections for the Convention that failed to deliver for him on Primary Day (except for the unions, which don’t appear to be so one-sided this year).
<
p>
Only Patrick really has a field operation for this campaign, but the smart guys behind Reilly and Gabrieli say that field doesn’t matter (but the smarter guys and gals behind them know that a good field operation can net 3 to 7 %). I guess we’ll see who’s right in September.
It’s not field or establishment. Patrick, Reilly and Gabrieli all know this. In order to win 50%+1 you need a good field team and inside endorsements. You need centrists and independents as well as a solid base.
<
p>
That’s what’s amazing, to me, about Deval’s campaign sofar. He’s created one of the largest grassroots organizations and garnered the support of 5 Congressmen that have been in the game for a long while. He’s gotten the Teacher’s association, no small union and a vast array of state reps, mayors, and other representatives. He’s getting donations from the grassroots, yes, but he’s also getting some heavy hitters in the Democratic Party, especially from the national African-American community.
<
p>
All candidates do this to some extent. I’m not about to underestimate Reilly’s field. He’s got some powerful unions behind him as well as Sal Dimasi and other insiders. And Chris Gabrieli started way later, so his field isn’t up to par–yet. But he does have millions to spend on ads and he’s working it on the ground level pretty hard. And as Reilly shoots himself in the foot, or as people learn more about Gabrieli, he’ll gain more supporters to do the grassroots organizing.
<
p>
Ultimately, all three are running different campaigns with different emphasis on grassroots, insider endorsements and machines and money on ads, but all candidates are working hard for all three. At this point, I’d say that Deval is doing the best with the grassroots and he’s certainly got a good number of insiders and party faithful. Moreover, his fundraising has set new records time and time again. But make no mistake: this is going to be a very, very close race unless something dramatic happens.
A teachers’ union endorsement. AFTMAss is a pipsqueak next to the Mass Teachers, which hasn’t endorsed.
<
p>
The thing that most impresses me about Deval’s campaign is how it retains this grassroots reputation despite the insider endorsements. The “establishment” loves this guy.
is not monolithic, and Deval plays on that very well. There are insiders who are progressive and insiders who are conservative. There are established progressives, like Newton Democrats for instance. They are there, waiting to be tapped. In some states that establishment is (much) smaller than others. Patrick, I think, does the best job of getting the outsider grassroots support and the progressive establishment. Nothing is ever all “insider” vs. “outsider” or “right” vs. “left.” That’s what’s so extraordinary about Patrick’s campaign: he’s working both the inside and outside. I think that will serve him well in the general.
The point of my comment, huntin’ dog metaphor aside, is not that Reilly isn’t doing as well as I expected — it’s that he’s not doing anywhere near what a truly popular and politically strong statewide official should be doing. Look next door in New York — is Spitzer second to a guy no one had heard of a year ago? Did a third guy in New York sense weakness and jump into the race late?
<
p>
Way less money than Gabrieli, much weaker organization than Patrick, fewer ideas than either of them. Real Democrat voters will mostly oppose Reilly’s tax cut — they think government ought to have enough resources to be able to do something. Independents won’t be excited about the only certifiable Beacon Hill insider on the Democratic primary ballot.
<
p>
Third place.
I realize you’re pro-Patrick, fine, but don’t tell me I’m not a real Democrat voter unless I agree with you.
I was distinguishing registered Democrats who vote in the primary from Independents who take a Democratic ballot. I also said that Democrats “mostly” would not gravitate to the tax-cut candidate — some no doubt will.
<
p>
I do not question whether you are a real Democrat, sabutai, and I agree with you about no circular firing squads. Reilly, Patrick, and Gabrieli are all light years ahead of Healey and none of us should lose sight of that.
You claim Reilly has “fewer ideas than either” of his opponents. You couldn’t be more wrong. Visit their websites and compare policy proposals. Reilly has more ideas and they are far more detailed than either Patrick or Gabrieli.
<
p>
And we’ll see on the income tax rollback issue, won’t we?
If you were around in 1998, you remember how Reilly went negative on Lois Pines. He’s the only Dem candidate videotaping his opponents, and I’m sure his opposition research is top notch (hey – he ain’t spending it on field ops). So batten down the hatches, and get ready for Reilly to spend $4 million on negative ads against Patrick and/or Gabrieli. It ain’t gonna be pretty.
because I moved here in ’97, but was so busy with the kid thing and the adjusting to a 3k mile move, that I don’t remember anything specific about this race.
<
p>
Could you (or anyone else) please elaborate on this Reilly v. Pines race.
<
p>
Thanks!
Reilly went negative pre-convention with a “Top 10 Soft on Crime Votes by Lois Pines” which pissed her off.
<
p>
She won the convention, but then imploded during the summer just as Reilly’s downtown friends got scared of her and helped him raise a ton of dough post-convention.
<
p>
Then she asked him to apologize for distorting her record during a debate – he didn’t – and she got killed for not backing off of that (being a woman didn’t help this, and it wasn’t nearly a Hillary-Lazio thing).
<
p>
The kicker was when she called Reilly’s supporters the “Old Irish Old Boys.” She was accused of being anti-Irish, pissed off “downtown”, and couldn’t recover.
<
p>
Finally, on election night, exit polls called it for her, but then the results came in and she’d lost (a la Kerry 04) 53-47.
<
p>
Yes, I was involved in that race tangentially, hence my stupid amount of knowledge about it.
<
p>
Bottom line: it was Pines’ race to lose and she lost it. She knew she was soft on crime but was able to get many police endorsements to counter Reilly’s prosecutor credentials (remember her ads marching with cops? hilarious, but also effective). Reilly got under her skin and rattled her and also convinced the business community that Pines would regulate them to death, so they gave him hundreds of thousands of dollars over the summer.
<
p>
It wasn’t a GOTV race, it was a TV race. There was also a Guv (Harshbarger, Donnelly and McGovern) and LG (Tolman and Kelly Gay) primary that year, as well as a Repub. Gov. primary between Cellucci and Malone.
<
p>
It was actually a pretty fun year.
It is true that the lower ballot positions – especially AG – were TV races, just like Treasurer in 2002. The most interesting result was how Kelly Gay almost beat Tolman. Harshbarger, though, had a real field operation in 1998, and a pretty good one too.
A different take on the 1998 AG race. It was a career prosecutor (Reilly) vs. a career pol (Pines). Pines had run for just about everything imaginebale, but without much success.
<
p>
DA and AG races almost always feature this contrast in resumes and the prosecutor almost always wins. Just look at how Jerry Leone cleared the field in Middlesex this year.
<
p>
When it comes to law enforcement, voters prefer a candidate with – surprise! – a law enforcement background.
Mav – That’s what the race should have been about. On paper, Pines was no competition for Reilly, but he went way negative on her, running ads about who contributed to her campaign and all kinds of other very nasty stuff. He should have won easily without going so negative. Heck, I even voted for him in 1998. I only mention this because it is clear that he will go negative this year as well, it’s only a question of time. It’s just his nature.
June 23 2002 HERALD POLL
O’Brien 31 %
Reich 19 %;
Birmingham 12 %
Grossman 6 %
Tolman 3 %
Undecided 28 %
<
p>
September 17 2002 ACTUAL RESULTS
O’Brien 31 %
Reich 25 %;
Birmingham 24 %
Tolman 18 %
<
p>
Hard to identify a trend from a single election.
<
p>
But since it’s fun to do, in that election the establishment candidate just held onto her 31%, while the newer faces divided up all the undecideds.
I wonder why the undecided right now is just 9%, while in 2002 it was 28% in June.
The poll methodology is highly suspect as previous people have commented on.
Not suspect — different. You can’t compare polls with different methodologies in this way.
Says that 36% of respondents have grad school and another 30% finished college. Anybody want to guess what the actually education of dem primary voters is statewide?
<
p>
This poll is no good.
The post-grad percentage is a little high, but the college grad percentage is actually too low. I wonder, though, if that’s because people misunderstood what “post-graduate” meant, and couldn’t ask the pollster to clarify.
<
p>
Here’s what the Statehouse News (traditional phone polling) had to say about dem primary voters in its May poll:
College: 40.6%
Post-Grad: 19.6%
<
p>
Add them up and you get 60%, and the MoE puts that within range of the SUSA poll.
<
p>
Look, I’m the first to admit when polls are crap, even when they favor my candidate. The Zogby Interactive poll, for example, is a complete farce. You can’t discount these polls, though, simply because they are automated-voice polls.
My point is we need to compare the breakdowns to the actual likley vote — not to another poll. Patrick won both the college grad and post grad groups and Reilly won the other two. Is the Dem Primary electorate REALLY so highly educated as that?
<
p>
Look, I don’t know the answer… but I don’t believe that either of these is an accurate sampling of the electorate.
Exit polls from 2002 general election have 58% college or higher. I would not be surprised if the Dem primary attracted a greater proportion than this.
<
p>
Anyway, this whole argument is pointless. You shouldn’t use these polls as predictors. They’re a snapshot in time. The trendlines are interesting, and you don’t run into these sorts of problems because the methodologies are kept constant.
Interesting that you posit Shannon O’Brien as the “establishment” candidate in 2002, as opposed to Tom Birmingham, then Senate President and the recipient of almost all the union endorsements and Mayor Menino’s backing, or Warren Tolman, who was the Lt. Gov. candidate in 1998, or Steve Grossman, who, as head of the Mass. Democratic Party was the true insiders’ insider. All I can say is that it sure didn’t seem like Shannon was the establishment candidate in 2002, when you had Joe DiNucci, Bill Galvin, and Tom Reilly trying to undermine her every step of the way. No woman in Massachusetts politics has ever been the “establishment” candidate, and it will take a real change of consciousness among the establishment for that to happen.
<
p>
And newer faces? Tolman, Reich, Birmingham, Grossman were newer faces? Please.