Susquehanna Polling released a poll showing Obama with a two point lead over Romney in Pennsylvania. Today on their blog they have a fascinating post explaining their methodology and why the makeup of the electorate is likely to be quite different from 2008:
Recently two polls conducted by our firm showing President Obama narrowly leading Mitt Romney by 1 point (48% to 47%, sponsored by the Republican State Committee of Penna.), and a second released by the Pittsburgh Tribune Review on Sunday, September 23 showing Obama leading Romney by 2 points (47%-45%). Both margins conflict with other surveys conducted recently including one by the Philadelphia Inquirer (Obama +11) and Muhlenberg College (Obama +9). Following are answers to questions about our survey methodology as well as our basis for predicting a close election.
Our vote model for gauging the number of interviews conducted with voters of different demographic groups (things like party affiliation, racial background and age range, etc.) is a blend of turnout models from both the 2008 and 2004 presidential elections, but leans more towards 2004 VTO and is predicated on the belief that turnout this November will not be anywhere near ’08 levels when 5.9 million votes were cast.
First, our ratio of interviews conducted with Republicans and Democrats in our recent polls (49D – 43R) gives Democrats a 6-point advantage based on the fact that Democrats outnumber Republicans in actual registration. However, this ratio is slightly more Republican based on both national and state polling showing that Republicans are more likely to vote than Democrats this year given high intensity among Republicans who strongly disapprove of the President’s job performance. Nonetheless, this +6 Democratic advantage is only one point less Democrat than the 7-point advantage these same exit polls gave Democrats in the 2008 presidential election. Besides, simply conducting more surveys with Democratic voters (as some have suggested) doesn’t necessarily translate into more votes for President Obama when you consider that Mitt Romney is winning Democratic-leaning counties in Western Pennsylvania by ten or more percentage points. Nonetheless, it is entirely appropriate to sample Republicans one or two points higher than in 2008 if you believe as we do that voter turnout this November will have little resemblance to the last presidential election.
Second, our ratio of younger to older voters reflects turnout that is likely to be slightly higher with older voters given the lack of enthusiasm from younger voters. In our surveys, 18-44 yr. olds make up 30% of all interviews and voters 45 years of age and older represent the remaining seventy percent. For instance, according to 2008 exit polls voter turnout among 18-29 year olds peaked at 18%, but national and state polling proves interest among younger voters down sharply this year due to higher unemployment with younger voters and college graduates in particular. So conducting approximately ten percent of surveys with 18-29 year olds is a reflection of this lower anticipated turnout among these less-enthusiastic voters. Besides, the fact that Obama backers have suggested that over sampling older voters skews results in favor of Mitt Romney is a striking revelation in a state like Pennsylvania known for having the 5th largest population of senior citizens in the country.
Go here to read the rest. Too many pollsters are merely plugging in the 2008 electorate and badly skewing their polls to Obama. Don’t believe them. Keep your eye on the Rasmussen and Gallup tracking polls to have the best feel for the race.
Please, Lord Jesus, may Romney – Ryan win! Thy will be done!
If they could just clean up Philadephia, PA would go to Romney. It’s the cities that throw the states to the Demoncrats. Every election is a sea of red from coast to coast except for the urban areas. Frustrating.
Good information to have! but is ok with me if they think they are ahead– maybe they’ll be less concerned about getting to the voting booth.
I’d like to know how soon early voting results will be reported? they could certainly skew things.. and is early voting a better outcome for liberals or for conservatives?
Disillusionment with politicians in the wake of the financial crisis seems to be producing low turnouts.
In the recent French presidential election, nearly one-fifth of the electorate did not bother to vote. At just over 80%, it was well down on the 88% turnout in 2007.
It does seem opinion polling is getting less and less reliable as time goes on. Then again, 35 years ago you might see published results from no more than four companies (Gallup, Harris, Roper, and Yankelovich – and only Gallup and Harris most of the time) and they conducted fewer surveys. Maybe it is just the sheer number of surveys manifesting problems that have always been present in market research.
Rasmussen has a precis of their methodology on their site. I think its premise is that it is not possible (or no longer possible) to produce a random sample with the means pollsters have to contact respondants, or that the sample you produce will differ systematically from the pool of people who actually vote. Therefore, their respondants are sorted into various demographic categories and then synthetic results are produced from weighting the subsamples according to census data, the results of other surveys, &c. Evidently, they are not able to contact cell phone users with random-digit dialing so they amend the results of their landline collections with those from online surveys. Heck knows if all of this kneading actually works.
Then again, one agency (Pew, I believe) produced a survey in which self-identified Democrats outnumbered Republican by 19 percentage points. There have been times in our history when that was true, but not since about 1977.
It is trite Art but true that polling is more an art than a science. In our very partisan age, where the parties are close to parity, I think party id is probably the most important factor in a poll, and if that is fouled up the poll is junk. The interesting thing is that even in polls where the partisan id is way out of whack, there are some where Romney does quite well. Linked below is an Ohio poll with a D-10 sample where Obama and Romney are neck and neck. If I were a strategist for Team Obama these type of polls would scare the devil out of me.
http://nicedeb.wordpress.com/2012/09/24/obama-and-romney-neck-in-neck-in-oh-poll-with-d10-sample/
PA unfortunately always seems to be one of those states where we have the potential to win over working-class Dems who are not cultural leftists (the “bitter clingers” as Bammer calls them) but we lose. dunno what this poll’s sampling was but there was another showing Obama with a double-digit lead so i’m not optimistic. it’d be great though.
it’s incredibly frustrating that Romney seems to have to play a mostly defensive game at this point, winning back states like Virginia without being able to make inroads into lean-blue states. i hate being the Concern guy cuz part of my dread is that in the event of a Romney loss the GOP will be pressured to become squishy-soft to appeal to [insert liberal voting bloc] in the future, so i hope they pull it off.
as far as partisan edge i think there’s somewhat of a Dem advantage. wasn’t 2008 D+8 or something? so less than that. Rasmussen uses D+2 i believe, dunno if that’s too low
nationally anyway, obviously states vary
“as far as partisan edge i think there’s somewhat of a Dem advantage. wasn’t 2008 D+8 or something?”
2008 was D-7; 2010 was parity. My guess is that the outcome this year will be parity to D-2. Republicans and conservatives are willing to crawl over broken glass to vote out Obama and that level of intensity matters in regard to turnout. The Democrats had the intensity in 2008 and they don’t have it in 2012.
I f you all follow the link, which takes you to a television interview broadcast about seven weeks ago, you see that they were interviewing Patrick Caddell (a pollster formerly in the employ of Jimmy Carter and Gary Hart) and one other pollster. Caddell said some of these surveys had samples so biased that publishing them was unprofessional. One has to be taken aback by that in the political atmosphere in which we live.
True Art. I do not contend that there is any grand conspiracy going on here, but most pollsters seem to be looking at 2008 turnout and blindly assuming the electorate will be the same this year. Leaving aside the repudiation of Obama at the polls in 2010 in the mid-terms, the enthusiasm gap between Republicans and Democrats and a bad economy, the pollsters are not taking into account the fact that registration for Dems is down in several key states:
http://www.nationalreview.com/campaign-spot/328401/democrats-advantage-voter-registration-slipping-key-states#
Most of the polls simply do not make sense, especially those positing that Obama is doing better in many states than he did in 2008.
I think Caddell worked for George McGovern as well, and at least one other candidate (Paul Tsongas, perhaps).
I am not sure what these news organizations are doing. I have participated in face-to-face small sample polls, which have the advantage of a high response rate (“robust”) and had good sampling frames, so you actually did not need to knead the data at all. The blogger “Dad29”, whom you see occasionally in precincts like this, says he works in marketing research and biased samples are a chronic problem.
One participant in these discussions at a Republican blog I look in on occasionally offered a hypothesis based on his own interaction with telephone pollsters: the people who make the calls have quotas to fulfill but are bedeviled by non-cooperation by the people called at random, so they have (on the QT) a set of people with a track-record of being patient and co-operative and call them to fill their quotas. That subset is non-representative.
Rommey is up 48-46 in the Rasmussen tracker today, and Rasmussen is using a D-2 projection. In 2008 Rasmussen saw the wave coming for Obama. He polls 15000 people every month on the question of party id. If Rasmussen isn’t seeing the electorate like 2008 it is because the Democrat wave just isn’t there this year.
There was a survey published on Monday for Ohio, showing a dead heat.
With a D +10 sample.
But for my money, the most egregious was the PPP poll for Missouri, right after the Akin magic uterus gaffe. It showed him in the lead against McCaskill. With R +9. A clearer case of attempting to influence a result (keeping Akin from dropping out) you will not see. Of course, it worked, Akin not being a particularly sharp Congressperson.
Rasmussen seems pretty sober, if with a slight inclination toward the Republicans. But he still gets it wrong, sometimes horribly–Sharron Angle in Nevada in 2010 being the most recent case in point.
“But for my money, the most egregious was the PPP poll for Missouri, right after the Akin magic uterus gaffe. It showed him in the lead against McCaskill. With R +9. A clearer case of attempting to influence a result (keeping Akin from dropping out) you will not see. Of course, it worked, Akin not being a particularly sharp Congressperson.”
PPP is completely in the tank for the Democrats. I agree with you Dale that this is the worst example of a phony poll this cycle.
PPP is completely in the tank for the Democrats.
Yes, but with an important caveat: they start taking serious, realistic polls within a month or so of an election. Every time.
They shake the pom-pons for Democrats for most of the cycle, then take the clown nose off and put real-world glasses on. Compare a PPP poll from three months out with their last survey taken before the vote: you’d swear body snatchers had radically transformed the electorate, if you looked only at the merry pranksters at Public Policy Polling.
“They shake the pom-pons for Democrats for most of the cycle, then take the clown nose off and put real-world glasses on. ”
Correct. I would like to say that I only see it with PPP but I have observed changes in numbers by other pollsters at the last minute, although not as radical as PPP.