Pollster.com

Articles and Analysis

 

POLL: SurveyUSA Pennsylvania Dems


SurveyUSA

Pennsylvania
Clinton 53, Obama 41

 

Comments
killias2:

Rasmussen has a poll today that has this race down to a 5-point difference. I trust SurveyUSA more, but I still think it is important to mention.

Either way, this race is narrowing, and I'm thinking it will continue to narrow. Obama won't win PA, but he can certainly bloody her here.

____________________

lemonfair:

ARG appears to be more like Survey USA. What are the differences in polling method etc to account for this? The polls have been very steady in the 51-39 range, so on the one hand the Rasmussen is a very large change, and on the other hand it's a surprise that Survey USA and ARG would show no change as a result of Obama getting in there and campaigning. Any thoughts?

____________________

the_real_truth:

It is interesting because Rassmussen had clinton up 10 points on 3/24 - similar to susa and arg. Now, it has been halved. Perhaps they are in tune with the shift - a bit before susa and arg? We will have to wait and see if the shift is real or just an outlier.

____________________

Thatcher:

lemonfair -

Actually - ARG (12% Clinton lead) was taken just before Obama started his bus tour and the Casey endorsement. MOE is 4%.

Survey USA (12% Clinton lead) was March 29 - 31 and has a MOE of 6.1%. This started the day after Casey endorsement and bus tour.

Rasmussen (5% Clinton lead) appears it was last night at MOE is 4%. The one with the most bus tour "under it's belt".

I'm not saying any one is the most accurate. Just making sure we know the time-line.

____________________

Thatcher:

Now, as I read my own comment - I have to laugh ... that puts the Survey USA poll in a virtual "dead heat" even with a 12% margin, because of the MOE! now THAT's reliability. Either Clinton and Obama could tie or Clinton could blow out Obama by 24 points!

Talk about covering the bases.

____________________

Thatcher:

***** CORRECTION *****

Sigh - I knew something didn't sound right -- so I hope people read this one before comment on my previous ones ...

(I was up late last night - is that an excuse?)

Survey USA MOE is 4.1%, I sincerely apologize for the error.

So, to correctly list:

Actually - ARG (12% Clinton lead) was taken just before Obama started his bus tour and the Casey endorsement. MOE is 4%.

Survey USA (12% Clinton lead) was March 29 - 31 and has a MOE of 4.1%. This started the day after Casey endorsement and bus tour.

Rasmussen (5% Clinton lead) appears it was last night at MOE is 4%. The one with the most bus tour "under it's belt".

____________________

Joseph E:

If you look at the numbers, Rasmussen has Obama 41% (almost the same - actually 1 point lower), but Clinton at 47% (6 points lower than SurveyUSA).

In the past, Mark has said that SurveyUSA pushes undecided voters harder to pick a candidate. It looks like Obama's supporters are solid for him at over 40%, while half of Clinton's lead is from folks who are leaning her way, but not willing to state a preference yet, if given the chance to be undecided.

SurveyUSA has tended to better, at least when it comes to the final week of polling before a primary. I think it is useful to see both methods right now. These two polls suggest that Obama could get within 5% if Clinton's weak supporters stay home, or even closer if many of them switch or he gets most of the undecided. But Clinton could certainly get 10 to 15% if there is no shift over the next three weeks, and could get up to 20% if she gets all of the undecided voters.

Clinton is still getting almost 50% in Rasmussen's poll, so an Obama victory looks very unlikely, unless there is a huge shift his way. If he got every "undecided" voter in Rasmussen's survey (he could also get her leaners to stay home, and have great turnout in his favor), he might get up to 52% in the best possible scenario.

Fortunately for him, if the polls continue this way, he only has to keep the loss in the single digits to claim a "moral victory." Clinton really needs to win by over 15%, better yet over 20%, to improve her chances of winning the nomination.

So, why no polls of Indiana? Isn't that the next state that "really" matters, according to the pundits? I'm baffled that SurveyUSA and Rasmussen haven't polled there yet.

____________________

Tracy Flick:

The Rasmussen poll reflects more the Bosnia story than anything else. Rasmussen says 47% of respondents say they are following that story quite closely.

Once the media drops the Bosnia story and moves on to something else I would expect Clinton's numbers to rebound, just as they have nationally as the story fades.

____________________

Eric and Mark, could I ask a big favor? I was wondering if we could have one handy guide to Rasmussen, Gallup, SurveyUSA, and anyone else we see regularly, noting differences in the following areas:

IVR vs human
Controlled demographics (gender, party ID, age, etc)
Methodology (forcing to make a choice, order)
Typical MOE (if they wind up the same!)

It's hard to keep track of all of these sometimes, and it would be nice to have whatever we can say in one place. I know that some is privileged info and some plain varies, but it would help.

Thanks to everyone else for beating me to saying everything I was thinking. Oh, and let's just wait a week (or three) to see what the next news cycle will bring! :O

____________________

killias2:

"ARG appears to be more like Survey USA. What are the differences in polling method etc to account for this? The polls have been very steady in the 51-39 range, so on the one hand the Rasmussen is a very large change, and on the other hand it's a surprise that Survey USA and ARG would show no change as a result of Obama getting in there and campaigning. Any thoughts?"

Actually, this shows a substantial shift since the last SurveyUSA poll. 3 weeks ago, they had a 19 point advantage for Clinton. Now, with some of Obama's campaign sinking in but not all, it's a 12 point advantage for Clinton.

I think the argument that SurveyUSA pushes undecideds to decide more is pretty strong here. If that's the case, Obama has the inkling of a chance to outdo expectations and win. If it's not the case, Obama still has a pretty good shot to cut Clinton's lead down to 7% or so, which is a moral victory of its own.

But we need more info before we can say any more.

____________________

lemonfair:

Thanks for all these comments. Do we have any idea whether these are only registered democrats being polled- or I should say, do we have any idea how the switched party registrants are tending?

____________________

killias2:

I read somewhere that they tended to be Obama supporters, but I have yet to see solid evidence of that.

Anyone with any evidence one way or the other?

____________________



Post a comment




Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.

MAP - US, AL, AK, AZ, AR, CA, CO, CT, DE, FL, GA, HI, ID, IL, IN, IA, KS, KY, LA, ME, MD, MA, MI, MN, MS, MO, MT, NE, NV, NH, NJ, NM, NY, NC, ND, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VT, VA, WA, WV, WI, WY, PR