Pollster.com

Articles and Analysis

 

POLL: PPP (D) Wisconsin


Public Policy Polling (D)

Wisconsin
Obama 53, Clinton 40... McCain 50, Huckabee 39, Paul 6

 

Comments
Gavin:

That seems a bit more on track with previous polls and really makes me scratch my head at the ARG poll released Feb. 16th.

Does anyone know where I can look to see how ARG breaks down their polling vs. PPP?

____________________

David :

It is quite obvious to everyone now that ARG is simply a farce when it comes to polling.

They have Obama up by 8 points in Texas and down significantly in Wisconsin. They, perhaps, mixed up the polls?

What a joke - these guys should just stop putting out their numbers. It is pretty sad to see them look so foolish, day after day.

How is ARG polling - asking 10 people and then extrapolating it out to 600?

____________________

David :

It is quite obvious to everyone now that ARG is simply a farce when it comes to polling.

They have Obama up by 8 points in Texas and down significantly in Wisconsin. They, perhaps, mixed up the polls?

What a joke - these guys should just stop putting out their numbers. It is pretty sad to see them look so foolish, day after day.

How is ARG polling - asking 10 people and then extrapolating it out to 600?

____________________

Tom W:

This absolutely seems more representative of the state (as a WI resident myself). Maybe ARG only polls registered democrats (of whom there are not a multitude, what with WI being an open primary state (why bother joining the party))

____________________

RS:

[I am copying my comment for the ARG poll, since I compare ARG with PPP. It appears applicable to the current PPP poll as well, except PPP now says "with conventional turn-out, Obama leads 47-44."]

Interestingly, ARG says Senator Clinton leads Senator Obama amongst Democrats 58-32, while PPP (Obama +11) earlier gave Obama a 46-44 lead! Both say Obama leads among independents/others - ARG 60-35 and PPP 63-25. Also, ARG says independents/others are 38% of their sample, while I estimate PPP puts this fraction at ~24%. But the huge Clinton lead among Democrats as per ARG gives her the lead in that poll.

On the other hand, PPP says "if conventional turn-out, Obama leads 46-44, but we estimate young/African-American voters should turn out much more, so he leads 50-39." Or words to that effect. As someone said earlier, weather could have an effect as well. Besides, Senator Clinton apparently is leaving Wisconsin a day earlier than scheduled, perhaps giving the impression that she's written Wisconsin off. Wonder how that will affect her support.

Echoing Mark B's earlier calls, I wish these polling firms would give more details on their sample and screening process!

____________________

FlyOnTheWall:

Eric & Mark:

Let me ask two question that've been bugging me about polling in Wisconsin. It seems, from the information they've released, that all the polls done in recent weeks are using a fairly standard likely-voter screen to identify likely Democratic voters. There are a few twists to the Wisconsin process that make that problematic. There is no partisan registration, so self-identification may be even more perilous than usual. And Wisconsin allows same-day registration - in the fall of 2006, 20% of voters registered at the polls - so identifying the likely voters may also be tough. But there's not much pollsters can do about those factors, other than offer a little more humility in their analyses.

What's been bothering me is that Wisconsin is a rare state that uses a single ballot. There are sixteen names on the thing, grouped by party, but every voter sees the same list. That's why when the locals poll, they tend to use more nuanced questions: the Badger Poll presents a list of all the major candidates to every respondent, just like the ballot; the Wisconsin Poll asks whether they intend to vote for one of the Democrats who's currently running or one of the Republicans who's currently running, instead of asking about general partisan preference or in which primary a voter intends to participate. But neither of those polls has been in the field in some months, and we're left with national polling operations, which seem to be using their standard instruments. But those force voters to choose between two primaries, when they'll actually be choosing among sixteen candidates. The upshot is that I suspect we're seeing some fairly skewed results out of Wisconsin, because the likely voter screens are winnowing out a fair number of potential swing voters.

Here's the other thing that's bothering me. Today's PPP poll again takes the jaw-dropping measure of rigging its results; their usual model produced a horse-race figure of 47-44%, but, they explain: "we expect Wisconsin will
see the same record levels of turnout from young voters and black voters that Obama has
been bringing out in other states. That should propel him to a double digit victory." I'm not sure where they're getting that. I haven't seen much evidence that young voters or blacks are, on the whole, accounting for a hugely larger share of the primary electorate in other states, even though their raw numbers are certainly up. And even if they were, I'd like to see PPP quantify both the basis for the adjustment and their current projection. At least implicitly, they seem to be saying that this is not based on observed trends within the sample itself (say, twice as many black and young voters identifying as "certain to vote" in the poll as the standard model would have predicted) but instead, on how they're massaging the data they've collected. That bothers me. One number they do give is 12% - that's the share of the electorate they expect black voters to comprise. This, despite the fact that in 2004 they were 6% (their proportion of the state's population) and in 1992 just 3%. And since overall turnout will certainly be up as well, PPP must be projecting that black turnout will triple or quadruple. I find that difficult to credit.

Am I out of line here? Or is there something wrong with PPP's approach?

____________________

tony:

ppp ranks even lower than arg on the susa report card..

____________________

Very interesting discussion, and I greatly appreciate the info on the unique ballot.

While PPP does appear to be cooking the numbers (even more than Zogby does) there is little doubt that traditional methods have proven very inadequate this primary season. I don't think it's just black voters, but young voters in general that they expect to turn out.

This is an important point because as the expectations are set Obama continues to beat them - which is big news going into the next round. It's also largely unfair and is definitely influencing perceptions. I think that it behooves pollsters to try to predict based on what they have seen as patterns in other states not only for their own reputation, but for how the data is ultimately used in the press.

I do believe that we should expect Obama by about 12 in Wisconsin based on my gut feel and read of my cheese-headed neighbors. But they have a tendency to surprise us all, as do the Packers. It's the level of surprise that gets the attention of the press corps, for better or worse. Pollsters probably don't like making the news on their own, and hence a lot of fudge based on what they've seen so far.

____________________

I think the what all these polls tell us is that nobody knows what will happen tomorrow.

____________________

illinoisindie:

all i have to say is when is survey USA going to put out the real wisconsin polll... most of these others are garbage

____________________

sam:

i think that no way obama is leading by 13 points so i looked at areport card from all pollsters an this ppp is ranked #26 out of 39 polling firms and error +/- 9%

____________________

C.S.Strowbridge:

If this poll is accurate, and that's a big if, then I can't see Hillary Clinton recovering in time for Texas. If Obama wins in Texas, I can't see Hillary Clinton winning this race.

However, even if Obama wins by 10 points, that would have the same effect and he could do that if he grabs the vast majority of the undecideds.

____________________

illinoisindie:

all i have to say is when is survey USA going to put out the real wisconsin poll... most of these others are garbage

____________________

richard pollara:

The thing that surprised me about the PPP Poll was the weighting of African American voters. Hispanics and Asians combine for about 5.5% of the population. That is just slightly less than the African American population. Yet the African American vote is weighted three times as heavily (12% vs 4%). What evidence do they have to support this assumption? It is a trend that I haven't seen in other states. In a prior post Eric Hare said he was predicting the race based on his gut feelings and his read of his Cheesehead neighbors. It seems as though PPP is doing the same thing.

____________________

tony:

rs/flyonthewall -- great analysis! thanks for checking out the sample.

____________________

Adam:

Wisconsin Census Facts:
http://quickfacts.census.gov/qfd/states/55000.html


White persons 90.0%
Black persons 6.0%

Given that nearly all blacks are Democrats, 12% seems quite reasonable. Presuming the remaining 4% are nearly all Democrats, then you could assume around 8%. But this is not the case; both Asians and Hispanics have a very significant chunk of their respective population that are Indep. or Republican.

____________________

Richard Pollara:

Yes, they appear to be going with a gut instinct and using some kind of math to get that answer. That was my point. I think they arrived at the right answer, but not by a method that can be considered credible. Or necessarily repeatable.

We'll just have to see.

____________________

G.G.:

Fly, you and I are on the same wall with such similar thoughts -- we really are at a loss without one of our fine local polling outfits on this. The nationals -- pollsters, media, etc. -- never quite get Wisconsin. (They probably operate entirely out of Madison, which is in Wisconsin but not of Wisconsin, dominated as it is by out of staters who avoid Milwaukee.)

In addition to the presumption of such a huge multiplier of so few AAs in Wisconsin, PPP also predicts that women will be 54% of Dem voters in Wisconsin -- but it long has been one of the leading states in women's turnout. Of course, media and pollsters attribute higher turnouts this time entirely to Obama, but actual numbers of women's turnouts in many states and more intelligent analysis (of Clinton's turnout of young single women, for example) belie that.

As for someone who said Clinton "abandoned" Wisconsin, that's also buying the national media meme. Then Obama abandoned Wisconsin on Thursday, and again yesterday, and again all day today, until he returns tonight -- after Clinton stops in four cities across the state today, after working Milwaukee all day yesterday despite the weather. And that garnered the top headlines in the largest paper in Wisconsin . . . which is online and available to anyone who actually looks at what locals see, where locals will vote tomorrow.

Keep in mind that the NY Times, for example, famously put up a front-page map some years ago that we still enjoy in Wisconsin. The NY Times put Minneapolis in Wisconsin! And keep in mind that the national media include many who only know Wisconsin from studying journalism at Madison, which takes you back to the top of this. . . .

____________________

Basic math: presume the primary vote splits 50/50 Democrat and Republican for total population. That means, if all races split evenly, you'd have 90% white, 6% black, 4% everyone else on both sides.

It's a known fact that African-Americans vote better than 90% Democrat.

We can keep things in balance by swapping the 6% share of blacks on the Republican side for a 6% share of whites on the Democrat side, leaving:

Republican: 96% white, 4% everyone else
Democrat: 84% white, 12% black, 4% everyone else

There's your 12% black turnout in the Democratic primary vote.

Bear in mind this is very simple indeed, and doesn't take into account imbalances in the everyone-else vote or in the overall primary vote. It's just a rough illustration.

____________________

RS:

I posted a comment, which is awaiting clearance... (maybe because I posted the exit poll links?)

But in essence: comparing Missouri's CNN exit polls for 2004 and 2008 Dem primaries, the African-American vote share did not increase much (15% to 17%). However, "young" voters - 18-44 - increased from 31% to 41%.
In 2004, 18-44 voters were 36% in Wisconsin, and African-Americans were 6%.

So PPP seems partly right - increasing the "young" voter fraction - but partly wrong, regarding the African-American vote.

____________________

GT:

ARG might be right. It's pretty obvious that since she made gains that PPP would come out of no where to say he's up so Obama Bots, don't be surprise if she wins tommorow or comes closer than you think.

____________________

Another Mike:

For FlyOnTheWall

There's good evidence that both youth (18-29) and Black share of the vote is significantly up this year compared to 2004. Not sure if it's up as much as this PPP poll projects, but it is not an insignificant change.

http://blog.washingtonpost.com/behind-the-numbers/2008/02/democratic_demographics_past_a.html

____________________

richard pollara:

Adam: In 2004 6% of the voters in the Democratic Primary were African American, 4% were Hispanic or Asian. PPP has doubled the African American number while leaving the Hispanic/Asian number the same. Perhaps they are right, but when you make a huge assumption like that it would be nice to have some empirical data to support it.

____________________

s.b.:

Read the poll. it says the actual results of the poll are 47-44, but they decided to tweek the results.

There is no reason for this. Standard turnouts are much more reliable measures. The turnout may differ, but you can't assume it will. That's not accuate or reliable polling, that's propaganda.


This poll shows a thee point lead for obama within the margin of error.

____________________

Breecer:

GT - don't get mad because Billary is going to lose......again.

It is time to concede.

____________________

Ferd:

Sorry s.b, the poll takes into account what has been happening lately - that is the last few primaries have shown very high turnout in the AA community as well as the youth vote. That's why Obama won by a greater margin than predicted.

It is a good predictor of what will happen tomorrow.

____________________

s.b.:

In the last PPP poll Feb, 11th, before cooking the numbers, Obama had a 4 point lead on Clinton, this poll shows a 3 point lead before number fudging. So Clinton has gone up one point, insignificant but worthy noting the comparison between the two polls, not between different polsters.

____________________

There will be no significant increase in African American turnout for this primary.:

As I have stated on other threads. There has been absolutely no significant increase in African American turnout in recent primaries or comparable primaries. This is a myth. In MAryland for example the black vote was down three percentage points over 2004. In Viriginia it was up 2% but the the black percentage of the poulation also increased in Viriginia.

There is no statistical reason to believe there will be any significant increase in African American turnout as a % of the vote in this primary.

____________________

Will :

Polls are supposed to be based on some degree of objectivity - so this really isn't a poll. But Obama will win!!

____________________

jr1886:

Let me chime in for a minute. The race will not be close and I am on record saying that here. Hillary will lose by double digits. Here's why, independents and young people will come out in large numbers and dwarfed all Hillary's modest turnout among blue-collar women

In Wisconsin, voters can register the same day and the Obama campaign is well aware of that so they have been targeting young people on college campus to come out in huge numbers and they will just check 2004 exit poll.

Again, don't be surprised tomorrow when Obama wins by double digits. Consider yourself warned.

Here's my prediction:

Barack Obama 57%

Hillary Clinton 43%

____________________

Ike:

A lot of this discussion reminds me of the discussion of the Des Moines Register poll earlier this winter. But if you recall, they had it right: lots of young people and independents, just as they predicted (in their likely voter model). So the question is: is Wisconsin more like Iowa, or more like (let us say) Massachusetts?

____________________

Cokie Roberts:

"Here's my prediction:

Barack Obama 57%

Hillary Clinton 43%"

What will you do if it's not double digits? I think you should have to send $10 to Hillary Clinton for every point under double-digits.

For instance, a 9 point Obama win means you send $10; 6 points Obama and you send $40.

I will say Obama +9.

____________________

Ike:

Actually, that's an interesting precedent, but it points in the opposite direction. The Des Moines Register polled, and generated results that were very different than anyone's model of likely turnout - but it had the guts to stand by the results actually generated in the poll, and results proved them right. PPP, on the other hand, polled and found results different than it expected. Instead of standing by them, it adjusted the results to meet its expectations of higher AA and youth turnout. We'll have to see about that.


____________________

jr1886:

I would not mind excerpt I don't want to give Billary my money to spend on parking lot and the incompetent Mark Penn. Other than that, it's a fine proposition if you would agree to give $10 to Obama campaign for very point higher than the +9 you predicted,lol.

____________________

Cokie Roberts:

"the incompetent Mark Penn."

He is incompetent. I could run in VA and get 30% of the vote.

I live in a split household, Clinton (me) /Obama (he). So, technically, I've given to his campaign (albeit begrudgingly).

____________________

jr1886:

Cokie Roberts,

Mark Penn is beyond incompetent. Remember this ridiculous statement from the chief strategist of the Hillary campaign:

Winning Democratic primaries is not a qualification or a sign of who can win the general election. If it were, every nominee would win because every nominee wins Democratic primaries

The "Where's the bounce" memo was the last straw for me, the same day I wrote a diary calling this guy the worst in the business. A lot of his statements show that this guy doesn't have a working brain to be chief strategist. One final note, he's the main reason Hillary is in this predicament right now. Early on, this guy used to obsess with polls to prove that Hillary was inevitable and a shoo-in, a risky strategy, for the nomination at the same time he failed to define Hillary in his own term to the American people.
Now, Hillary's grasping at straws to save her bid for the nomination. She can still redeem herself but it won't happen until she fired this miserable loser i.e. Mark Penn.

So do we have a deal on the proposition above?lol

____________________

Gregor:

ARG is awesome! They have not been right about a single result this cycle! Do you realize how useful that is? Go ARG!

____________________

Progny:

I work in political polling and this is the first cycle I've ever heard of PPP.

Now, I'm not saying their data should be tossed out of hand, but quick glance at their website raised a number of red flags for me -- namely, a short client list, exclusive use of IVR methodology, and this...

"Dean Debnam, founder and President of Public Policy Polling (PPP), has over 30 years experience in creating and growing businesses."

As in, not polling, political or otherwise.

That said, ARG is still the worst for my money.

____________________

Brian:

Obama 58%
Clinton 42%

I think Obama winning by 16% in Wisconsin is a good bet. The Obama camp will roll in the Madison and Milwaukee regions. We can finally say goodnight/goodnight to the Clintons' on March 4th, 2008.

The African American turnout will make up between 12%-15% of the Dem voting turnout. And in this open primary you can register to vote on election day. It will be cold in Wisconsin, but Obama supporters will be feeling the sweet ole warm Obama love.

____________________

The Jesus and Mary Chain:

Wednesday February 20, 200--Arlington, VA:

While disappointed to have not done better in Wisconsin, campaign strategist Howard Wolfson said, "Our goal was to keep it close in WI, and we achieved that. Now we move on to our main objectives of OH and TX where Senator Clinton is making strong inroads on her way to victory."

____________________

jr1886:

I assume these are the same people that tel us the race would be over on Super Tuesday. They are in for another surprise tomorrow.

____________________

Perf:

Can someone tell whose polls have been more accurate in the previous elections?

____________________

jr1886:
____________________

rowjimmy:

Another new poll...

ARG shows a 16 swing in one day. LOL

http://americanresearchgroup.com/


If, after all the media talking up this being a close race, it turns out to a 15 point loss, I think that's the final straw for Clinton.

____________________

Michael:

American Research Group
2/17-18/08 600 LV
Clinton 42 Obama 52 Undecided 5

LOL :-)

____________________

If any of you are in WI or TX on primary day/s, could you take pics with your phone and upload them using the widget on this link? http://my.wauw.fm/Brazza

We are Obama supporters in Europe and would like to publish photos/comments of election events, capturing the spirit of the the contest outside of the international media newscasts we receive.

Thanks in advance

____________________

Henry:

lol. It's as if ARG is hedging their bets with these latest numbers.

Here's the deal, by my math:

If Obama wins WI today by 10 points (will probably be at least that), and wins Hawaii w/ 2/3, and goes on to lose in TX by 10 points (more likely to be a tie), and lose Ohio by 20 points (likely to be less than that), and basically split RI and VT, he comes out after March 4 still ahead significantly.

In other words, unless we see a big upset in Clinton's favor at this point, Obama is still going to be ahead significantly even after March 4, which Clinton has been banking everything on.

Extending it past there it becomes less clear.. but assuming Clinton decides to do further harm to the Democratic Party and doesn't concede after Mar 4 as she should if she has any honor left in her, if the battle continues all the way to Pennsylvania April 22 (at which point Obama will almost certainly be even further ahead having won more in Wyoming and Mississippi), even if she wins Penns. by 20 points she will STILL be losing in the delegate count, and Obama will be that much closer to clinching.

The Clinton campaign, even as amazingly incompetent as they are, must realize this and must be feeling pretty desperate. This is why we're already hearing talk of breaking the rules they themselves and everyone else voted for and counting the Florida and Michigan delegates (even though Obama didn't run/wasn't on the ballot there), and also why we're hearing so much talk about Superdelegates.

The Clinton campaign knows they're going to continue to lose and will lose barring the above dirty tricks succeeding, or Obama making some huge blunder.

Will be fun to watch things pan out.

____________________

Caden:

Good post Henry.

Working against Hillary is the proportional system of Democrats, that is Obama still gets a large number of delegates even if he loses Texas, Ohio and Penn by a 55-45 margin. She simply can not catch unless a few remaining states are "upset" wins. Very unlikely. In addition, Puerto Rico's influential governor already endorsed Obama. They have around 63 delegates, which Obama will get the majority of.

____________________

Richard Pollara:

According to CNN early exit polls AA vote at 8% and Latino at 4%. This is at odds with the 12% PPP assumption.

____________________

Richard Pollara:

Sorry shouldhave included this in last post: Male 43% Female 57% (CNN exit polls).

____________________

jr1886:

Cokie Roberts,

I nailed the results and Brian did as well. So please consider dropping a few bucks to Obama,lol.

My prediction was 57% to 43% in favor of Obama.

Actual results, 58%-41% Obama.

So here you have it Cokie Roberts, consider sending $80 bucks for Obama yayyyyyyyyyyy.

http://my.barackobama.com/page/outreach/view/main/Jr1886

____________________

ao:

great work on the poll. i don't think anyone else was remotely this accurate. we here in wisconsin are feeling really good today, knowing michelle obama's our next first lady.

____________________

jr1886:

Thank you ao.Without your support over in WI, I am sure I would be way off. Let's get to work to end this thing on March 4th.

____________________

Eric:

Great poll.

____________________



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