Channel 7/Suffolk poll out yesterday. sco, as always, has the slice-‘n’-dice.
Among Dems, Patrick’s in the lead:
Deval Patrick 31%
Tom Reilly 25%
Chris Gabrieli 22%
Undecided 21%
And check out the cool tracking graph from Suffolk.
Tom Reilly said on Greater Boston tonight that he’s not interested in polls right now. So he doesn’t care about the convention endorsement, doesn’t care about polls … You know, I’m beginning to wonder what Tom Reilly is interested in. Seriously, all snark aside, it’s one thing to position oneself as the “moderate choice”, but it’s another thing to go out and get voters — you know, create some buzz, actively gain folks’ confidence. The Reilly campaign seems to feel that if he camps out on that putative middle ground, folks will come to him eventually. Awfully passive for what’s clearly going to be a hard-fought primary.
Can some of the pro-Reilly folks fill me in on the strategy? What am I missing?
Thanks for posting it.
<
p>
A quick glance at the trendlines show this:
<
p>
Reilly trending DOWN
Patrick trending UP
Gabrieli trending UP
<
p>
Other opinions?
Here is my take on a couple of trends, one as noted above, the other re finances. The first graph doesn’t answer the objection noted below about the timeline being distorted, but it might make it more clear what the trends are, because I’ve included the “undecided” count, tallied (by implication) elsewhere by another poster.
<
p>
As noted by lolo, the Gabber is picking up voters from Reilly, not from Patrick — Deval’s lead over Chris actually grew from 6% last month to 9% this month.
<
p>
<
p>
I agree with the observation that winning begets winning in this biz. People love to back a winner. And Patrick has been working hard, with much success, looking more and more like a winner. The Convention win was pooh-poohed in some corners, even mockingly called the “Convention Curse” — but guess what? It got the attention of the press, and it turned a “Deval who?” campaign into a serious one. Not that that’s all there is to it, of course. Many of us have been bustin’ our butts going door to door, talking to friends and neighbors, holding house parties, etc.; the point is we haven’t been waiting for the voters to come to us.
<
p>
I think the “winner” image is having a very desirable feedback effect — as Patrick climbs in the polls, he raises more money, which makes him look more like a winner, which raises his standing in the polls, which…
<
p>
<IMG src="http://www.mfw.us/politics/polls/Monthly-Results.gif"
Saying the convention is not definitive in predicting a primary victory is true. Saying it doesn’t matter at all is false.
Someone had to say it. It’s like the crazy aunt/uncle in the attic. What is he thinking?? And I say that as a Patrick supporter in the context that Patrick tanks and Reilly’s the guy. Reilly’s passivity may be intended to masquerade as confidence, but, at some point, one has to ask is he campaigning for people’s support [read: vote] or isn’t he?
<
p>
My fear: September 19th rolls around and Reilly’s the guy–by dint, probably, of some force we cannot now quantify or imagine– and he’s passively sitting around saying, figuratively if not literally, “If I run, they will vote for me.” Not.
Um…pray for rain on Primary Day?
Check out the graph’s first hash on the y-axis, Feb 2005. There seems to be a stray Gabrieli point with data value 4%. I wonder why they didn’t connect it with the others. Sure it would go “through” two data holes, but so what? To leave it out there in the bottom left without drawing attention to it is bad presentation of data.
<
p>
Another interesting piece of data: the percent of people who chose a candidate:
Feb 05 48
Jun 05 36
Feb 06 69
Mar 06 63
Apr 06 64
May 06 70
Jun 06 78
<
p>
As that number climbs, it will be harder and harder to move, since instead of gaining the support from an uncommitted voter, you’ll have to wrest support from a voter who was currently supporting someone else.
<
p>
And finally, one more critique of these graphs, which really aren’t very good. The y axis uses points that are equadistant. Trouble is, the polls weren’t equadistant in time. The first two data points should be far apart, and the last five should be quite close. Divide the y axis into the total time (Jun 06 – Feb 05) and then put the data points on their appropriate timeline slot. This gives you a much better sense of the trend over time, which is what we’re really interested in.
But the most interesting trend is that Gab has apparently been pulling voters away from Reilly. It will be interesting to see whether that trend will continue. Sco is right to ask about what the Reilly camp has in store to stop the slide.
There’s no reference to any voter conversion data. Both Gabrieli’s and Partick’s increasing numbers come from a combination of shifts in undecideds and Reilly supporters. No way of demonstrating your claim from the data presented.
<
p>
Indeed, starting from the Feb-2006 data point, the total % of poll respondents stating a preference were 69, 63, 64, 70, and 78. Suggests a shift in undecideds as well as some movement from Reilly supporters.
<
p>
But you can’t tell with any specificity from where, to where.
lolo is clearly correct
<
p>
undecideds going down
Reilly going down
Gabrieli going up
Patrick going up more than Gabrieli
<
p>
ergo, Gabrieli is not taking votes away from Patrick
Reilly is not taking votes from anywhere
<
p>
former Reilly and undecided votes are going for both Gabrieli and Patrick, but more so for Patrick
<
p>
what’s unclear about that?
<
p>
now, if you want to argue that the poll is “meaningless” or “unrepresentative” or some such, you are entitled to do so — but if you take the results at face value, there is only one inescapable conclusion: Reilly is sputtering…
<
p>
I might point out, btw, that polls are always meaningless unless your candidate is doing well — so this poll is obviously VERY meaningful! but don’t worry — I’m sure there’ll be many flutters over the next 12 weeks…
lolo is not clearly correct.
<
p>
One hypothesis is that lolo is correct — Patrick is gaining undecideds, and Gabs is gaining undecideds and former Reilly supporters.
<
p>
Here’s another hypothesis:
Patrick is gaining former Reilly supporters. Gabs is pulilng only formerly undecideds.
<
p>
Here’s a third hypothesis:
Gabs is pulling votes from former Patrick supporters. Patrick is pulling (even more) votes from undecideds and former Reilly supporters.
<
p>
You might argue that some of these hypotheses are more plausible than others, but the scientific conclusion is only that all three are plausbable, and the data present offers no way to confirm with (if any) of the above three hypotheses are correct. The only way to know for sure would be to increase poll size, and then track the same sample (same 2000 people) over the time-series polls, so that you could see what each of the 2000 people did, and then extrapolate that onto the MA population.
to be able to look at a graph and figure out that Reilly’s numbers are going down almost proporionately to Gab’s numbers going up while Deval’s numbers are shooting higher. Sheesh. Michael Wilcox is an economist. He seems to get it too. Maybe we should both take lessons from you so that we can analyze data more correctly.
Lolorb … the confusion is that the Globe omitted a line showing “Undecideds.” If they were to include that it’d be a lot clearer.
<
p>
The reason this is important is that I’d like to know where new Patrick and Gabreili supporters are coming from. Are they fleeing Reilly for Gabreili? Or are new decideds flocking to Patrick?
<
p>
All hypotheses presented are plausible within the constraints of the actual changes in percentage. But I contend we cannot tell from the data precisely what is happening.
<
p>
Do the math. Here are the differences between the May and June 2006 data points:
<
p>
total decideds: 70% increasing to 78%…add 8% points
Reilly going down…give up 10% points
Gabrieli going up…add 7% points
Patrick going up…add 11% points
<
p>
For all we know, 100% of the breaking undecideds (8%) could have fallen into the Patrick column, accounting for 8 of the 11-point Patrick swing, with Reilly’s loss of 10% split 3% points for Patrick and and 7% for Gabreili.
<
p>
Or, 7 of the 8-points of new “decideds” could have broken for Gab, and the remaining 1-point and ALL of Reilly’s former supporters (10%) could have broken for Patrick.
<
p>
We can’t tell from the data. So it’s impossible to say whether it’s Gab or Patrick picking up former Reilly supporters (in excess of the obvious 1-point minimum for Patrick.)
Upon first glance at these graphs, I too thought that the non-linear plotting of the data on the ‘time’ axis was simply misleading. After further review of the tape, this referee has changed his mind. I believe that a useful x-axis for poll data should be days-until-election on a semi-log plot, and these plots aren’t too far from that mark. Public interest in an election does not grow linearly with time, it is more of an exponential type function or a power law. These types of functions should be plotted on a log scale if you are trying to see a trend in the data. Another ‘independent’ variable that we might consider using on the x-axis is simply the fraction of undecideds – I’ll bet that will make an interesting plot. Good luck with that, Dr. Wilcox!
You’re asking a lot of the Globe. Perhaps in the future they can use a logarithmic scale. I call to see if they can plot the x-axis using a first derivative.
I can see this is going to become a full-time job!
<
p>
btw, did you notice your “Lies” post has disappeared from the airwaves? (As did “shack”‘s post about rural voters…) What gives? Didn’t pass the NSA screen? Taken away by a black helicopter?
Tax rates? That was this week’s story. We are moving on, Michael. If you did want to visit the “Lies” post I think you can always get there via my bio page. (The NGIA doesn’t yet have that location on their maps!)
Seriously, I’m thinking we should dig up some Kerry/Bush poll data from 2004 to see what it looks like in a few different plot types. When does the public start ‘thinking’ about who they are going to vote for? Or, when do the undecideds actually decide? We might get a better plot from Gore/Bush 2000 since it was slightly less divisive.
<
p>They’re on the front page now.
As best I can tell the Reilly strategy is this:
<
p>
(1) We will get 15% at the convention no matter what, so don’t spend unnecessary time or resources there, only to be embarassed by not winning the endorsement/preventing anyone else from getting endorsed;
<
p>
(2) No one pays attention in June/July, so we need to be out there enough to remind people we’re there, and keep our people in line;
<
p>
(3) We’ve set the spending limit at $1.5 million (maybe 1.2, it is early and the cobwebs are still clearing out) and TV spots are going to be big for us;
<
p>
(4) The 6 week push: in the final 6 weeks of the campaign (August and Everything After), we start going up with TV and reminding people why they’ve voted for Tom Reilly twice for Attorney General and reminding the most populous county in the state why he was their DA twice. The ads feature law and order and the grandkids, the humble beginnings in Springfield, the still humble life in Watertown. Reilly’s voters (moderates, blue-collar, maybe some old Reagan Democrats)aren’t too sure about this guy who their over-enthusiastic neighbor from two streets over stopped by the house to tell us about (Patrick). That guy is always making the town meeting run long with the petitions about the Patriot Act. I think Tom Reilly understands us. We vote for Reilly.
<
p>
That is my best guess. I can’t speak to the buzz of the campaign as I am not in the bay state. From afar it does look increasingly good for Patrick as more voters start to make decisions, I am just not entirely sure how firm those decisions are. Assuming Gabrieli spends his dough, a lot of minds could change. And if Gabs goes for Patrick’s throat, Reilly may be the only one left standing.
<
p>
I don’t think you’re really missing anything Charley. I assume there is something less-passive planned, but I am not privy to the inner-workings. As things stand now, it certainly seems like there is just a lot of hanging around.
Good guess.
… and plausible. But even so, if one respects Patrick and Gabrieli as worthy foes (which they are), why would you (for instance) cede the caucuses, the convention and a lot of media afterglow to Patrick? Why wouldn’t you want to crush that campaign like a bug to begin with, or at least try to dampen the enthusiasm?
<
p>
Put it this way: If the initial convention delegate counts had been 40%-40%-20% (Deval-Tom-Gab), It would still be a big story for Patrick, but I doubt he’d be in the lead right now, at least not by as much.
<
p>
Winning begets winning. No, nothing counts as much now as it will in August and September, but I don’t see Reilly laying the groundwork for a successful primary season.
My understanding is that Reilly did not “cede” the caucuses, but he did not think he was going to lose so many of them. His people knew that Patrick was organizing a good ground game, but Reilly himself, from what I’ve been told, didn’t think he had anything to worry about.
I started supporting Patrick a long time ago, but when I first started telling friends and family the universal answer was, “Who?”
<
p>
More importantly, later on, those who had heard of him (in my experience) assumed he would be an also-ran, a guy on the liberal end of things just to have the debate. I imagine that sort of thinking permeated to the Reilly campaign.
<
p>
Reilly’s been shooting himself in the foot allowing Patrick to seize the opportunity.
<
p>
Anyway, long comment short, it’s been a long, long ride and it’s only now starting.
Sorry for being late in replying Charley, but I agree that this was a pretty big tactical snafu. In some movie, there is a line that applies (I think maybe the Godfather, sorry for being cliched): “If you’re gonna hit a guy, you knock him down so he can’t get up again.”
<
p>
If I were the Reilly campaign, I would have mounted a big-time effort at the caucuses. Imagine how much more validity his message has if he can say “I am the more moderate Dem. (he wouldn’t say this, but essentially insert his talking points here) and that appeals to a super-majority of my party as well as independents.” It would be great publicity for a campaign to be able to say: “Well we dominated the caucuses, so we are clearly the party’s choice, this Patrick guy might get on the ballot, but he’s this year’s Robert Reich.” Reilly could have essentially ignored him. I think this underscores your point of winning begetting winning. And the more Reilly looked dominant, the easier time he would have had raising $$$.
“That guy is always making the town meeting run long with the petitions about the Patriot Act.”
<
p>
Yes, yes. Since it’s relying so heavily on word-of-mouth, the Patrick campaign needs to be careful with its brand.
…but I would have to disagree with you on this point:
<
p>
The nature of a grassroots campaign is that it produces much firmer support. Think about it — if a voter responds to a pollster that they favor Gabrieli, but that just happens to be because they are a Red Sox fan and saw his TV ads during every game they watched that week, what will happen when the other two candidates start advertising? That voter will probably remember the last ad they saw.
<
p>
Contrast that with the typical Patrick supporter, who has gotten to know Deval personally at a rally, or heard him on a conference call at a friend’s house party, or has written the campaign a check (or all of these things). How likely is that person to be influenced by a Gabber attack ad?
<
p>
All of this is why I think Chris and Tom will be fighting it out for second place, and at this point, I wouldn’t want to have to bet my November mortgage payment on which one will be the runner-up. But it seems to me that they will both be increasingly defensive, and (sadly) more negative, as they realize they are battling over an increasingly diminishing pool of votes.
<
p>
How’s that for unbiased analysis?
But if it is, current poll standings are not too meaningful. They may show trends, general attitudes. But I’d wait until 30 days before the primary to pay close attention.
<
p>
Grassroots campaigns may work fine in primaries with low turnout. But they’re never as effective in general elections as having boatloads of money to spend on advertising as voting day nears.
<
p>
That’s why I have to think that Gabrieli, thanks to unlimited funds, has the ability to snatch 5 or 10 points (more?) in the closing weeks.
<
p>
Just my guess that money still has clout in a primary.
Michael – You’re mixing apples and oranges. You can’t compare a voter supporting Gabrieli in a poll with a “typical Patrick supporter.” Both Gabrieli and Reilly have supporters just like us, just not as many. But we typical Patrick supporters are only a small part of Patrick’s support as measured by polls.
<
p>
What really scares me is Reilly’s history of going negative. It’s gonna be hard for him to go negative on Gabrieli after he almost chose Gabrieli to be his Lt. Gov. running mate. My sense is that he’s gonna go after Patrick and Gabrieli together – comparison ads – Reilly’s for cutting taxes and the other guys aren’t. But I wouldn’t put it past Reilly to use the approach Ed King used against Frank Hatch when King had TV commercials of Hatch’s home in Prides Crossing – look for Reilly to do something like that against Patrick and the house in Western Mass. that is like a fetish to Reilly.
<
p>
Also, anyone else notice bumper stickers when driving around the state? I’m used to seeing Patrick stickers all over, and Reilly stickers in downtown Boston, but I don’t think I’ve seen one for Gabrieli yet.
I rarely see stickers, at least for the gubernatorial contest, but when I do they are for Patrick. In Watertown I’ve seen some “Firefighters for Reilly” stickers, though.
MFW, sorry for not getting to this sooner …
<
p>
As usual, good points all around. I would say you’re right in that the very nature of a grass roots campaign is that the bond between candidate and voter, and voter and other voters is stronger. In part this is because the voter isn’t just a consumer, buying into an ad or a slogan, but an active participant in the campaign. I don’t think the Patrick campaign is alone in having dedicated activists, I just think it focuses on them more, and does its best to include/pretend to include them in really running the campaign (I tend to think it does include them, but this could be debated).
<
p>
My only real point of contention is this: do you think that the polling results reflect the strength of the grass roots network, or feelings derived from free media along with grass roots contact? I guess what I am asking is that surely you don’t believe that 30-odd percent of voters who support Patrick have all met him, do you? I guess it is possible, but I imagine that some number of those are liking what they’ve heard in the debates and through the news. Certainly some of the support is from people who have been convinced by friends/neighbors etc. but I don’t know how much of it is.
<
p>
That said, you’re very correct (if one can be more or less correct) that grassroots campaigns develop firm support. And while that is useful for caucuses, visibilities, etc. is it the whole show on primary day? I think the Patrick campaign is pointed in the right direction, but support an inch wide and a mile deep won’t necessarily win things. Granted, his support isn’t an inch wide, but I’d say that Reilly’s isn’t just an inch deep either. And the way he has positioned himself, I think he is more likely to be able to find new sources of support.
<
p>
Someone else posted that the Reilly people are praying for rain on primary day. I’d say the Patrick people should. You know you’ll turn out your voters. But then again, some good liberals, er progressives that the politics of hope attracts wouldn’t want rain because everyone deserves their say right? Just kidding you a bit, hope this made some degree of sense (long week).
<
p>