I just saw the new Boston Globe poll, showing a slim 45-43 lead for Scott Brown. It’s scary, and should be taken seriously, but what’s up? Almost all other recent polls show Warren with a lead just outside the margin of error. It’s not the top line number that makes me question the poll, it’s the secondary question about the Presidential race:
The poll shows that Obama leads Romney 52 percent to 38 percent, a substantial margin, but down significantly from last month, when Obama held a 27 point lead over Romney in Massachusetts.
That just does not make sense. Ten percent of voters are undecided or going third party? Obama’s number is only a bit better here than it is in a battleground state like Wisconsin? Obama’s lead has been cut in half since September?
I guess we will have to wait and see the crosstabs to see what could be going on.
That said, though, this poll at the very least confirms that this race is incredibly tight and Warren needs all the help we can muster over the next week.
The Globe Poll has been weird throughout. I find their sudden shift to be unlikely and, more to the point, it comes when nothing else is going on or the campaign does anything different. As has been said, one poll means nothing, but the Globe has been an outlier among many of the polls with its grossly high undecided rate. Also, hasn’t UNH’s polls been called into question more generally.
All that said, unless Brown shaves his head, joins a monastery and drops out of the race, we should not miss a beat and treat every minute we’ve got like this race is as close as this poll suggests.
I think he’d *gain* votes if he became a monk.
today. The ads been around for a little while I think, at least a few days, but it is incredibly negative and misleading. It says that Elizabeth Warren wants to increase payroll taxes? She does? I know the payroll tax cut is temporary, but anyway, he has spun that one to high heaven, because I don’t think she has even made a statement on that. Anyway, I have seen no Elizabeth Warren ads. Granted, I am only watching TV intermittently, but people could be tuned in today because of the storm, and because they are home. She needs to rebut that ad. The campaign needs to watch it and rebut every single negative attack, because I am worried it is reaching too many voters on this dreadful day. I can’t find the ad on YouTube.
in about 10 minutes on the radio this morning.
I am not sure about the Globe/UNH poll. Haven’t seen the cross-tabs but they have been weird this cycle. Last time they had almost 20% undecided. This poll suggests Obama’s down to 52 percent in Mass. and has over 10 percent undecided in the Presidential race. I don’t buy any of that.
No panic but everyone make sure to vote, make sure all Warren supporters you know vote, and if you can, volunteer to GOTV!
As I see it, this should be a classic turnout-driven election. Presidential election year. Democrats primed to max turnout. What equivalent turnout resources does the Brown camp have? I suspect that the survey probe questions for “likely voters” are faulty.
GOTV
This ad has been playing ALL DAY on NECN.
That is annoying.
Also, support for a balanced budget amendment should be enough to be disqualifying. It’s just a crazy and dangerous policy. But, most voters probably see an ad like that and think that Brown is for a balanced budget and Warren is against, without understanding the amendment issue.
Brown hits every low note with this one.
BTW, here in CT, senate ads have been playing on TV non-stop all day — Murphy and McMahon both. I think they have made the calculation that running ads during a hurricane is no biggie.
In 2010, with 9 days left to election day, Martha Coakley was 15 points AHEAD of Scott Brown in the polls…the rest is history. Hope we are not seeing a repeat.
Ironically, that poll was the Boston Globe, which suggests it was actually wrong and overstated Coakley’s edge. The polls shifted sharply to Brown immediately after that, which tells me they are overstated the “favored” candidate as they may have now.
With consensus growing that Brown’s negative attacks have harmed him more than her, until we see more polling, this just does not indicate that sharp of a swing in his direction, especially given this poll’s shoddy history.
Everything about it is different. Presidential race vs. off-season special election, candidate with huge grassroots team and drawing big crowds vs. candidate who had neither, united Democratic party vs. fractured Democratic party. Coakley didn’t act like she wanted it and didn’t build the alliances that lead to victory.
Ignore the polls. Stay focused. Knock on doors. Make phone calls. Hold signs. Raise money. Together, we will win.
Brown’s betting the same blend of anti-tax, anti-government, anti-affirmative action, anti-government-funding-for-abortion-or-contraception, anti-“liberal agenda” that the Boston Herald, Barbara Anderson, and Grover Norquist have been pounding into our ears for years will work again. Remember, Grover grew up in Weston, and remember his strategy for this election — from his speech at CPAC: http://www.youtube.com/watch?v=6wYYX0mZsQA
doing a larger sample. 1,067 would only decrease the margin of error by 1%, but I’d like to see the cross-tabs. In last month’s poll by these guys, they had a 23% response rate. It was also an automated poll that weighted its responses.
I’d much rather see a few news agencies “throw in” and drive down the margin of error. Get it down to the 2s. It takes a lot more polling, and this isn’t just about getting the top line “right” — it’s about making the cross tabs useful, which means that each of those cross tabs [gender, age, urban-ness of community, employment status, education level, home-ownership, etc.] have a usable margin of error.
I’m a professor of statistics. “Margin of error” always assumes you have a genuinely random sample and everyone tells the truth to a neutrally-worded question. It’s ridiculously easy for biases to creep in–for example, most people hang up their phone when a pollster calls. Are the ones who stick around to talk voting with the same patter as those who hung up? If only a handful of under-30 people talk to you, do you assume those 5 people are representative of the entire under-30 population and give then huge weight, or do you assume that very few under-30s will vote.
As your sample size gets large, the theretical Margin of Error shriks, but these biases–which involves enormous amounts of guesswork to try to reduce–stick around. So that all you accomplish is that you look foolish when your “1% MoE poll” is off by 4.
I get the idea that there are (at least) two MoEs at play… the theoretical, the biases in responses, and maybe others.
But when the MoE is reported at 4.4%, that too is the theoretical MoE. The additional biases aren’t part of that calculation either.
I’m not so interested in getting the top-line result theoretical MoE down to 1%. I’m interested in getting the cross-tabs’ MoE down to something sensible. That the top line MoE would go down to 3%, 2% or whatever is an artifact, not the goal.
Yes? No?
If Obama goes from a 27 point lead to a 15 point lead in the matter of a couple of weeks, when his support has fallen off no more than 3-5 points anywhere else in the country at most, it’s pretty hard to lend credibility to a poll that suddenly has Brown up 2 instead of down 5 like the rest of the country. Heck, even Rasmussen has Warren up 5, and we know about their 3 point GOP tilt.
This time of the election cycle, they pull their methodology in line with the rest so as not too look like an outlier in the last few weeks of the election. So generally the bias is there, but several weeks before it disappears. So this is incorrect in a manner of speaking.
They do this so that schmucks will still hire them in the next election cycle, cuz they can say, SEE? Our poll was pretty close to the final result!
Meanwhile, they try to be, essentially, a push poll in that by having a bias for most of the election cycle, it SEEMS like the Republican is doing better than he is, and that might make people think, well, if other people like him/her…
Or at least, that’s my best guess as to why Ras does what it does.
the Obama/Romney numbers seem way off in this poll. Also odd, as Paul McMorrow pointed out on Twitter, is that according to this poll, 49% of registered Republicans have never heard of Charlie Baker. ??