CQ TODAY ONLINE NEWS
June 27, 2008 – 12:17 a.m.
The Wild Differences in The Polls, Explained
By Andrew Satter, CQ Staff
Video: Pew Research Center editor Richard Morin explains discrepancies in different presidential polls
If you’ve been watching Presidential preference polls over the past week, you might feel a bit whipsawed.
For the second straight day, Gallup’s daily tracking poll Thursday has John McCain and Barack Obama tied. Both candidates dropped a point from yesterday’s tracking poll, down to 44 percent. The margin of error is +/- 2 percentage points.
The Rasmussen daily tracking poll in the same period of time has shown a 3-7 point gap between the two candidates.
Neither of the daily tracking polls square with two polls from that show Barack Obama holding a stunning double-digit lead. A Los Angeles Times/Bloomberg poll showed Obama with a 12 point lead over McCain, and a Newsweek poll had Obama leading by 15 points, if Ralph Nader and Bob Barr were included in the mix.
Why the difference? Were the Newsweek and L.A. Times biased in favor of Obama? Do the Rasmussen and Gallup pollsters favor McCain? Or maybe the public is wildly changing its views, daily.
All unlikely, says Richard Morin, a senior editor at the Pew Research Center. In an interview with CQ Politics, he said the discrepancy is probably a result of the Newsweek and L.A. Times/Bloomberg polls over-representing Democrats.
“When I look at those results, I know something is going on,” said Morin.
“The first place that I look when I see these discrepancies, I look for the percentage of Republicans, Democrats and Independents in the sample. We know that the best predictor of how someone is going to vote is their party ID.
“Both the L.A. Times/Bloomberg and the Newsweek polls have (too) large percentage of Democrats and a (too) small percentage of Republicans.”
While there are indeed more people who identify themselves as Democrats than Republicans in the country, Morin says the other polls, including Gallup, are more in line with the actual disparity than the Bloomberg or Newsweek polls.
“Interestingly enough,” Morin said, “if you do the math and apply the proper percentages to the L.A. Times/Bloomberg and the Newsweek findings, you find that their results change dramatically.”
In fact, Morin says, if the two polls that show Obama winning by a large margin were to modify their findings using the same percentage of Democrats and Republicans as other polls, Obama’s lead would come down to somewhere between a toss-up and a small, single digit lead for Obama.
The Wild Differences in The Polls, Explained
Morin cautioned that neither organization oversampled Democrats intentionally. It’s a tricky business, getting exactly the right mix that mirrors the entire population.
“Embedded in the operations of polling are subtle differences that result in skews — and they can be either Republican skews or Democratic skews.”
Andrew Romano writes in his Newsweek blog, ‘Stumper:’ “The problem here is that unlike race, age and gender, party ID is fluid--and even extreme swings might reflect actual changes in the mood of the electorate. ‘This is a canard,’ says NEWSWEEK polling maestro Larry Hugick when asked about the McCain memo [which charged that the party ratios were wrong]. ‘Both parties do it. But ID isn’t a fixed property. In fact, it’s associated with the candidates. It’s been proven that as a candidate goes up in the polls, so does his party. Same when a candidate goes down.’”
As the season wears on, there are likely to be plenty of differences among the slew of surveys. Pollsters commonly caution that each poll is “just a snapshot in time.” And the focus has to be just right.
Comments
It is not uncommon of the media to run biased polls such as the LATimes report. Their white guilt makes them prime candidates to support any minority, no matter how unqualified. The willful distortion of reality is justified, since the outcome of Obama appearing more qualified is their goal. In their minds, the ends justifies the means! Once again, showing that color does matter. Martin Luther King Jr, must be rolling over in his grave at this straightforward racism of the media and democrats.
I think we will see this a lot this election season - wildly different polls, where the chief differences are the underlying assumptions about party membership and turnout. My guess is that the real number is somewhere between Gallup (which I think is being too conservative here - there has genuinely been a switch in partisan preference) and Newsweek (maybe not that much of a shift). The truth of the matter, though, is that none of these pollsters have any idea what is going on with party preference - this is an entirely different world from 2000-2006. So, take these polls with two grains of salt.
Newsweek has become part of the Obama campaign and should not be considered at all. They have zero objectivity and appear to be actively supporting him. Like a French judge in the Olympics one must ignore their score as it has no value.
I agree with Charles, above, that these things can skew both ways. I like the site electoralvote.com (i'm not affiliated in any way, just a fan) because the site does poll averaging, which tones these outliers out, and helps give a better picture.
I don't really look at the polling anymore. After the disaster that is New Hampshire I couldn't really trust them. I get my hopes up that somethings going to happen and then I see that disintegrate because of incorrect polling. If you look at the prediction markets I think they're allot more interesting. They don't show what people "want" but what people "think." Wanting something to happen and thinking something will happen are very different. For example a play money prediction markets like Hubdub has Obama with a 70% chance of winning, Intrade a real money market also somewhat mirrors Hubdub's chances for Obama. Instead of relying on peoples feelings I thinks it's better to rely on the "wisdom of the crowds" which usually turns out to be right. I'd like to see a CQ analysis of prediction markets and whether they are more reliable than polls...?
Which begs the contextual question: Why are Newseek and the LATimes working so hard to convince Obama's supporters his numbers are strong enough to win the general election without putting Clinton on the ticket?
After New Hampshire I just can't trust polling. I do doubt that Obama has a 15 point lead but I don't think the election is as close as Gallup portrays it. Either way I don't see why we depend on polling when there's a better alternative, prediction markets. If you go to prediction markets like Hubdub.com or Intrade.com you'll see predictors currently give Obama a 70% chance of winning. Prediction markets rely on the analytical "wisdom of the crowds" while polling relies more on what people "want to happen." With prediction markets users are predicting because they want the payoff, so there is more of a incentive to be right. I'd like to see a CQ article explaining how useful prediction markets can be, in comparison to trusting polls?
The video contradicts Morin's own thesis. He says, ""Both the L.A. Times/Bloomberg and the Newsweek polls have (too) large percentage of Democrats and a (too) small percentage of Republicans." And then says that his own organization has a sample where there is a 9% point difference between Dems and Republicans. But if you look at the sample size for the Newsweek poll he's criticizing (the one they show in the video), you'll see that Democrats are 36% and Republicans are 26% -- only one point off of what he claims should be correct. Hard to believe that one point difference leads to a "significant" change when he adjusts the sample to his own liking. I don't know whose poll is right, but it certainly seems that Morin's criticism is wrong.
Chris, thanks for your comment. I should clarify that Mr. Morin was correct in his assessment that the Newsweek and L.A. Times polls overaccounted for Democrats. You are correct that the difference between those who identified as Dems and Reps was only 10 percentage points, but the poll also asked the people who identified as Independents which party ID they leaned. A vast majority of these people said they lean Democrat, thus creating a large gap, somewhere between 17 and 19 percentage points between those who identify or lean Democratic with those who identify or lean Republican. As Mr. Morin states, the actual number is somewhere closer to 9 percentage points. Sorry for the confusion.
Wouldn't the more accurate sample be the percentage represented by either party of the total number registered? In truth, using a 50-50 proportion is not the way thing are in this real world.
If anyone is showing bias here it is Morin, who seems to assume that Newsweek and the LA Times are the ones who are putting "too many" Democrats in their sample mix. But the fact is, it is Gallup and others who are re-weighting their samples to reflect what they think is the "real" party ID division in this country. Newsweek and the Times are using exactly what their ramdom samples tshowed -- which is a massive tilt towards the Democrats. Is that a real trend? Will it last through November? Beats me. But I think it is Gallup et. al. who should be required to justify their arbitrary manipulation of the data to match what THEY think is the "right" party balance.
Even if the Democrats are slightly overrepresented, you're forgetting that at least some are Clinton supporters, so it may well be that he is getting most of his biuynce from Independents, which would hurt McCain a lot, because the Clinton supporters are only about half way towards getting behind Obama. It's likely as the campaign wears on and becomes a referendum on the Bush years that that 70-80% of the public who think the country is headed in the wrong direction will make the two party split in this poll look generous to the republicans. Does anyone think housing prices or gasoline prices or anything in the domestic economy will look any better by November? Will McCain say anything to make that part of the public who voted Bush back in think he'll do a better job than Bush? Don't think so. PeteR
It was a random sample. So the insinuation is that the pollsters overrepresented Democrats with their list of randomly picked numbers? I think that before people start throwing these accusations around they should research and how pollsters get their sample.
Could it be that 49% of the electoral base now identify themselves as Democrats, while only 37% identify themselves as Republicans, which is why you find "more Democrats than Republicans" in the sample? That puts the margin of error at 12pts, doesn't it?
It is no accident that skewed polls favor Obama 100% of the time. The media has been in the tank for him the entire election process, and can share credit for his "appointment" as the Democratic nominee. Voters must realize that we are in this alone, and if we wish to be governed by the candidate of OUR choice, we must work twice as hard as the media has in foisting Obama on the public. In 2000 and 2004 the media promoted and protected Bush, and we all know how that came out. Now they are promoting and protecting Obama. I doubt that this is a coincidence. I am a lifelong Democrat, but I will be voting for John McCain in November.
Every poll makes an assumption, a screen or filter, to rejigger its raw numbers into the turnout they expect. Rasmussen's is usually too optimistic (from a liberal pov), Gallup's too pessimistic (same), etc. Reverse if your flavors are different, but the point is, if Obama's weird new turnout map delivers, he is elected, and if not, not. All the rest is just ad buys, annoying voters at home, and GOTV. Everyone, go, think about summer.
Fred, Anyone who claims to be a life long Dem and is voting for McSame this fall is either nuts or a closet racist. Obama has been chewed to pieces by the media. Just the "Wright" mess alone generated hours and hours of wasted media attention. Meanwhile teflon John skates. Between Keating, his trophy wife, and cheating on the campaign funding process, McCain is no McSaint. We can't afford 4 more years.
I already knew what Mr. Morin explained, becasue I read Protein Wisdom. Maybe Mr. Morin does too.
I think the article is generally true. However, one point not highlighted that in random sampling, sampling error is a function of the number of samples to be taken, among other parameters. The greater the samples, the lesser the accuracy. Using the samples alone, Gallup is more correct as it uses greater number, about 4000+. Of course, these polls are not true reflections. Because, there is locational weight difference where the sample is taken - the issue of electoral votes per state. If to be more accurate, stratified sampling should be used. More accurately, do random sampling on a per state basis - which gives us more reliable snapshots. But why an objective article be discussed in a partisan way? Common people! stop all your dogmatism to each of your party. Your divisiveness is what pushing this country down. Your individual ideology (open and hidden) really makes everyone in this country stupid. And stop the money talks and trading! You are all shameless. As if presidency should be dictated by buying or by militancy or by activism. This very thing is why the world hates America. And MSM, please state facts and educate the readers of real facts about both Obama and McCain. Stop using the "lying politicians" (flip floppers and inconsistent) as the basis of your facts. Endorsement and muddling of issues are for politicians. MSM roles is to give "objective data" and information from records and documents for each issue raised by lying politicians. Preferably, triangulate the verification of infos about these politicians MSM makes the Americans more ignorant of the importance of Presidency. MSM blamed all the Presidents, both Bush and Cllinton - now are trashes of the society. If the public negative perception of these presidents are actually true - then MSM should be the one to blame. Because you spread wrong information and signals to the people. Common! stop the hypocrisy of the MSM. Return all the money obtained by MSM to the politicians which are also obtained from the people (lobbyists are people too). MSM is now acting like lobbyists in reverse (bought by politicians instead). Just for this once, MSM should have a sense of patriotism and gratefulness to the people by giving them correct, balanced and unabridged information for each of these candidate. Review both McCain and Obama's past and expose them to people. This election has caused so much division already.
POST A COMMENT
Oops! The following errors must be addressed: