A new CBS4/Survey USA poll of 479 likely Democratic primary voters taken over the last few days shows the same lineup (Patrick, Gabrieli, Reilly) as the previous one, but the differences are small, and the margin of error covers most of them.
Here are the results of the latest poll, which was conducted Monday 7/31 through Wednesday 8/2. The results from the last SUSA poll, conducted 7/9 through 7/11 on 417 likely Dem primary voters, are in parentheses.
Patrick: 35% (37%)
Gabrieli: 30% (27%)
Reilly: 27% (26%)
Undecided: 8% (10%)
MOE: +/- 4.6% (4.9%)
The movement from the last poll strikes me as so small, and so well within the margin of error, as to be insignificant: bottom line message seems to be “no change.” Which suggests that, rather than the CA/Tastrophe giving Reilly a boost (which I had thought it might), it seems to have put the campaign into a state of suspended animation. No effect from the Big Dig, no discernible effect from TV ads. That strikes me as good news for Patrick, since it’s better to be frozen in first place than not.
UPDATE: The fact that everyone (so far) who has commented on this post thinks I’m wrong – they think that the movement from the last poll to this one is significant – has encouraged me to try to dig a little more deeply into what the margins of error in polls like these really mean. It may take me a bit of time to slog through the statistical concepts involved. This helpful post by Kevin Drum (from 2004), first called to our attention by joeltpatterson, is a good start, though it raises more questions than it answers for someone like myself who really doesn’t know that much about statistics. If you’ve got useful pointers, drop ’em in the comments; meanwhile, I’ll be doing my homework and will get back to you.
FURTHER UPDATE: Sure enough, I was wrong. See my discussion here.
The movement from the last poll strikes me as so small, and so well within the margin of error, as to be insignificant: bottom line message seems to be “no change.”
<
p>
I disagree. I think that this clearly shows that the race is tightening. Before, the gap was 10%. Now its 5%. I’m not arguing that Gabs has passed Deval, but he’s clearly making inroads.
<
p>
It’s also worth noting that the percent of undecided has fallen by a fifth — there’s only eight more percent to go around — not really enough for Reilly and probably not enough for Gabs either.
<
p>
It’s not great news for Gabs or Reilly, and its not really bad news for Patrick. Still, I wouldn’t call it “no change”.
It may be that the ads and Big Dig stopped the bleeding for Reilly. He has been trending down in SUSA since March. It might be that the ads are keeping him from falling further.
Gotta disagree on the “no change” diagnosis. Patrick’s lead over Gabrieli went from 10 points to 5 points, from well outside the margin of error to just on the edge of it. Reilly’s deficit went from 11 points to 8. Patrick has gone from front-runner to having a small edge in what is looking more and more like a dead heat. By the time Patrick gets on the air, there’s a good chance he and at least Gabrieli will be polling statistically even.
<
p>
Warning – sheer speculation follows: Good news for Patrick supporters. If the race is even with, say, three weeks to go, and Patrick is only starting his ads, then the resource gap is largely leveled. So it comes down to three weeks of whose ads connect better, who does best in debates and interviews, etc. Then there’s Patrick’s field organization to GOTV for an extra couple of points (beware those who claim it’s worth much more than that).
<
p>
But the tightness of the race and the attendant horse-race coverage could even the score for Gabrieli and Reilly, by bringing more voters to the primary for reasons other than the GOTV. In this poll, Gabrieli leads among independents, 38-32 (DP) -25 (TR), and conservatives, 52-23-17. Reilly leads among moderates, 36-31 (CG) -27 (DP). I think people who are drawn to vote by the closeness of the race, and not by a specific candidate’s organization, are more likely to fall into these categories. I also hope some Democrats will listen to the idea that these numbers position Gabrieli best in the general election.
<
p>
I’m looking forward to some high-stakes debates.
Much like the general, I think at this point it will come down to GOTV. I agree it will be aphoto finish, and the debates will ge big.
<
p>
I’m also surprised that the three candidates have kept from pulling the negative ads trigger so far. But smart of them to do so…you do it with a week left, there isn’t time to be punished. You take up all the air in the room with your negatives, and by the time we’re trying to punish the one who started it, primary’s over.
Gabrieli leads among independents, 38-32 (DP) -25 (TR), and conservatives, 52-23-17. Reilly leads among moderates, 36-31 (CG) -27 (DP)
<
p>
Who do you think is going to turn out for a Democratic primary of a non-presidential year? Conservatives and independents? I think even if the polls show a “statistical dead heat” going in, that’s very good news for Patrick.
<
p>
BTW, I believe Patrick will be very well positioned against Kerry Healey and Christy Mihos in the general election. The last Republican candidate this weak was Paul Cellucci, who squeaked by only because the Democratic “establishment” repeatedly stabbed their own nominee in the back. They won’t make that mistake again – or they won’t get away with it in this environment, to put it another way.
This is going to be one tight race which is all about GOTV in the end.
As Publius’ post focused on: Patrick is not on TV while Chris and Tom are. I, like the others here, have to disagree with David who said this poll represented, no change. Chris was 10 points down, now five points. Reilly gains 3 points. True, there are no champagne bottles popping at Reilly and Gabrielis headquarters and no ones jumping out the window at Devals. But a lot of smiles are at the former and stomachs are churning at the latter.
<
p>
Can Patrick afford to let Reilly and Gabrieli have another 3 or 4 weeks of unanswered TV? Can Patrickliterallyafford to go up on TV? Last info I read/heard is Patrick has $1.5 million in the bank. That buys himself an OK three week TV buy after Labor Day.
<
p>
What well find out real soon is how much personal resources Deval really does have. Can he afford, with an estimated $25,000 a month personal overhead, to part with at least $1 million to stem the lose of another 5 point lose to Chris and even ahve Reilly in striking distance? And if he does have an extra million, can he afford to folk over another $2 million for the week after the primary to go up against Healeys anticipated attacks.
<
p>
As I said over on Publius postDevals game plan has always been retail/grassroots. Hes been spending like a drunken sailor on this strategy. If hes not going to buy TV, he needs to crank up this field effort and maybe these numbers will scare the field into action. Last May, I question the strategy of spending early. Among the things I questioned was spending $45,000 for the April rally at Faneuil Hall, if I may quote from the post: Now I’m sure many supporters will contend it was worth every penny. Well, we’ll see how you feel the weekend before the Primary when you realize that money could buy the campaign another 15 spots on the various local news outlets. Those 15 extra spots could have allowed the campaign to reach thousands of undecided primary voters in a very close election…of well, but it was a great rally.
<
p>
Hey, the Deval camp invested in field and they knew that who keep them off TV until after Labor Dayunless Deval is richer than I think he isthat is the plan they made and that is the plan they have to stick with.
<
p>
Besides, so far the plan has worked. So far.
I made these comments on the poll.
<
p>
Point is that Deval is still solidly in the lead. That doesn’t mean it couldn’t change — we’ve all been around too long to take anything for granted — but as a snapshot it is a solid result for the Patrick campaign.
<
p>
“Frank” is sceptical of the Patrick strategy of doing grassroots organizing instead of media buys. We’ll see. My hope is that we make believers out of everyone because that will truly change the way politics is done here in Massachusetts and maybe in the nation. That is my fondest hope.
<
p>
Meanwhile, all eyes on CT!
Kevin Drum’s article is helpful and elaborates what I was going to say before I saw the update: You can’t just add together margins of error.
<
p>
To elaborate, the “margin of error” is a range within which you have a certain amount of confidence (usually 95% certainty) that the actual number lies. So if Patrick’s polling at 35% and the MoE is 4.6%, there is a 95% chance that Patrick’s “real” number is between 30.4% and 39.6%. So far, so simple.
<
p>
Statistically, though, you can’t add together (or subtract) two numbers and apply the same MoE. In this example, Patrick’s range is, as I said, 30.4% to 39.6%; Gabrieli’s range is 25.4% to 34.6%. Not all values of the range are equally likely, though, so simple math doesn’t compute.
<
p>
Fortunately, though, Kevin Drum provides a link to a simple spreadsheet which will do the necessary heavy lifting for us. We see that in reality the odds of Patrick being ahead of Gabrieli are 91.31%, and of being ahead of Reilly are 98.73%. Here you can, in fact, multiply the odds together and see that, if the election were held today, it is 90.15% likely that Patrick would win.
<
p>
Pretty good odds, but not insurmountable, and notably less than 95% confidence. So statistically, from a 95% confidence level, it is not a certainty; but I think anybody here would take a bet they had a 90% chance of winning.