“Markey’s lead over Gomez shrinks” claime the Boston Globe. “A new poll from Suffolk University suggests that we could have another close special election for a Senate seat in Massachusetts on our hands,” hypes the Washington Post.
But is the latest Suffolk University poll truly signal? Or is it noise? Is this the start of another 2010-like trend? Or do these media reports show that journalists haven’t learned that polls move up and down during the course of a campaign season and such fluctuations aren’t necessarily statistically significant?
If you haven’t seen it yet, the Suffolk poll released today shows Markey ahead of Gomez by 7 points, 48-41. Yes, that’s a major decline from the prior Suffolk poll — but, I don’t think any of us really believed that previous poll showing Markey up by 17. Bad comparison. How does it stack up with the other independent polls of late? Well, today’s poll remains within margin-of-error range of every other independent poll since the beginning of May.
Nate Silver hasn’t weighed in on this race since May 6, but at that time Silver said that Republicans “can win Congressional elections in blue states if just about everything goes right. . . . However these instances are rare. Mr. Gomez is capable of winning, but his roughly five-point deficit to Mr. Markey in the polls now could also prove to be a high-water mark.” Gomez’s deficit has still exceeded 5 points in every independent poll since.
So right now, I’m calling this latest change in Suffolk polling results noise, not signal. I’d need to see more data with a considerably more significant downward trend before I’m ready to declare signal.
stomv says
but the bars should be spaced by the time between polls (and probably centered on the middle day of the poll). The first four polls are all clumped in early May, then we had one poll mid May, then four polls clumped early June. By providing the spacing, it emphasizes that
(a) the Suffolk poll on 5/4-5/7 is a clear outlier,
(b) the Emerson poll on 5/20-5/22 isn’t closely related with the NECollege poll 10 days later,
(c) Markey appears on a slight downswing in the clump of June 1-June 4 polls, continued in the most recent Suffolk poll.
Still, the shift is slight, and it’s hard to know if it’s measuring a tightening of the race or just random noise.
oceandreams says
but more as ordered discreet events, yet time matters. Real Clear Politics’ average of the most recent independent polls is probably as good a “slice in time” as any, though, and that’s showing Markey +9.8.
What gets me nuts, though, is the same media that happily declares any poll results within margin of error “a statistical dead heat” nevertheless reports changes from poll to poll that are also within margin of error range as significant. They need to pick one and be consistent.
This Suffolk poll was the first one taken after news of some but not all of the federal surveillance. If any Democrat’s support will be affected by popularity of the president of the same party, that would be something I”d like to look at in upcoming data.
hlpeary says
Markey 46%, Gomez 39%, new WBUR/MassInc.; Markey was 8-pt lead early in their poll in May. http://www.wbur.org/2013/06/11/gomez-markey-wbur-senate-poll
Think back on Coakley- Brown race. Coakley was sinking in the last 2 weeks and they brought all the DC poohbahs in to help, including the President, and that made it worse because Brown was able to play off the “the Beltway Boys” …this race could get tighter yet…
oceandreams says
Sorry but a change of 1 point in a poll with a margin of error of +/- 4.4 is not statistically significant. I’m willing to show concern at a change of 3 or a lead that edges close to margin of error, but this is neither. Even the writeup you link to says that Gomez is struggling to chip away at Markey’s “small but consistent lead.”
Key here is consistent. This reminds me of last year’s presidential race that so many pundits were trying to portray as in doubt, even as President Obama held small but consistent (and statistically significant) leads in most key swing states, Congressman Markey continues to hold consistent and statistically significant leads in every single independent poll.
Would I be happier if the lead were in double digits? Yes. Is turnout of Markey voters critical for this race? Beyond a doubt, especially in a special election where the electorate is tough to forecast. But is this data showing a significant change in Markey’s lead from the last WBUR poll? No.
One more note: Markey’s lead is an identical +7 in this WBUR poll with or without undecideds who say they lean toward one candidate or the other. Which means an awful lot of undecideds have to break Gomez’s way or people have to change their minds in the next two weeks for this to change.
danfromwaltham says
How weak of a candidate, a 37 year political veteran need I remind you, Ed Markey is. As bad as Gabriel Gomez is, 46% is all Markey can poll? That’s where Ed is with two weeks to go, in a state that Obama won with 61%? What does this say about a candidate?
Any doubt, Scott Brown would have wiped the floor, with Ed Markey?
Problem for Gomez, people like me have no interest in the race. I find it more intersting walking around my yard, picking up after my dogs, than listening to those Gomez and Markey.
So, Mr. Ed will win in a pathetic voter turnout by simply being a Democrat, in an overwhelming Democratic state, and not by being Ed Markey. Wow, what a victory!
merrimackguy says
Why hasn’t Markey “turned it up”? Doesn’t look like he can. Kennedy always did despite being a perennial lock.
Gomez previous political experience- lost a selectman’s race. He should be at 30% (pretty much the anti-Democrat base), not 40%.
Issues: Markey has a couple (but not much that are “his”), and Gomez just dodges them. Probably what the public is looking for (so as not to strain their brain during the summer months).
Turnout/ground game: This is the complete wildcard. Republican activists are tired and entered this disheartened. Gomez was not the candidate of the true believers who tend to work more. Republican and Republican leaning voters do tend to show up. We’ll see how the Democratic machine works in the warm weather, as Democratic voters are often weather challenged.
stomv says
If he’s very likely to win because the race is low key and that maximizes all of his built in advantages (name recognition, Democratic party label), why on earth would he even want to “turn it up” and draw attention to the race?
The goal isn’t to earn as many votes as possible. The goal is to maximize the probability of earning more votes than Gomez. Methinks that Markey’s quiet focus on ground game is doing just that.
merrimackguy says
You don’t think he needs to inspire, lead, etc.? To use the campaign to convince people that there are bigger ideas out there that need airing than just “elect the guy who’s not the other guy.” ?
Why did Sen Kennedy always (always!) act like he was asking/working for your vote? It made people feel better about having him around, like the country was a better place because he was on the job. Markey’s no Kennedy, but he could at least act like he’s got a pulse.
I think (not that dbag expression methinks) that politicians should try to do a bit more than get 50% +1, but maybe that’s too much to expect. I now get why elected officials only expect to represent the people that vote for them, not the population as a whole.
stomv says
as an elected official, your job is to inspire, lead, etc. Sure, you could argue that Markey is both candidate and elected official right now, but we’re talking about his role as a candidate.
And that one should behave within the confines of both the law and ethics should go without writing; should.
merrimackguy says
Maybe he was thinking like you. As candidate Obama he would just say whatever it took to win- that was his job.
He never intended to do those things he talked about as a candidate.
Again I’m learning something. Candidates are only saying and doing what they need to win. Don’t expect anything else.
I’m not sure what this means, but the ethics of winning gives a candidate some pretty broad latitude.
merrimackguy says
as we have seen in the past. If so, just indicate and I’ll move one.
oceandreams says
We got kind of spoiled here with Ted Kennedy as our senator, but he was hardly typical. This is John Kerry’s unexpired term. What exactly was John Kerry doing to “turn it up” in recent elections where he had comfortable leads and “no-name” opponents? I’ve seen Ed Markey on the local campaign trail a heck of a lot more than I ever saw John Kerry.
merrimackguy says
try to emulating the best? John Kerry couldn’t even “turn it up” to beat one of the worst presidents in modern history. Ask anyone who was on that campaign. “Flawed Candidate” is the most cited reason for that loss.
oceandreams says
If you look at the polling done for the 2010 special election at Real Clear Politics, there was an independent poll of likely voters in October, almost 3 months before the special election; the next polls of likely (as opposed to registered) voters weren’t until early January, just 2 weeks before the vote. Polls couldn’t have shown us a steady decline in Coakley’s support because none of the independent firms were polling the race seriously throughout, thinking it was a foregone conclusion.
There has been a much steadier stream of independent polling data this time around.