Rasmussen vs Gallup – what’s going on?

April 25, 2012
By

Up until now I have viewed Gallup and Rasmussen as equally reliable in their polling. Generally they have been closer to each other than either was to the other pollsters that publish public polling results. Recently that has changed.

Rasmussen and Gallup are now both producing daily head to head tracking polls on Romney and Obama. Over pretty much the same period Rasmussen shows Romney up 5 and Gallup shows Obama up 7, a 12% difference well outside of either polls MOE. It’s also interesting that, over the last 7-10 days, the Gallup poll has shown a lot more movement than the Rasmussen poll. Pretty strange why are they different? It appears that the reason is that the two polls are using very different assumptions about voter demographics.

Rasmussen is kind enough to publish their demographic assumptions. Rasmussen conducts a three month moving average of the party demographics that each month produces data for a sample set of 45000 voters. So as of the three months from January – March of 2012 their sampling indicated the following party affiliation for the electorate: 36.1% Republican, 32.8% Democrat and 31.1% unaffiliated. This is by far the most comprehensive analysis of voter demographics by party affiliation of any polling organization.

Gallup is not as forthcoming about their demographic assumptions, but it is possible to work backwards from their reported support by party and come up with a very close ( less than +-0.5%) estimate of their sampling assumptions. The problem with working this backwards is that rounding of the actual reported sample induces some error preventing a perfect fit and at the edges of the sample sizes you have the same error margins. However one can come close enough that we can establish the sample sizes well within the MOE of the poll. Working Gallup’s results backwards they are using a weighted sample that is about 27% Republican, 33.5% Democrat, and 39.5% Independent.

If you take the party specific results of the Gallup poll and apply the demographic assumptions used by Rasmussen you get 44.8% for Obama, 47.7% for Romney and 7.5% undecided. Given the poll’s respective MOEs the difference between the adjusted Gallup results and those for Rasmussen is not statistically significant. What is significant is that Obama goes from up 7% (outside that poll’s MOE) to down 3% (inside that poll’s MOE). Further, when Rasmussen’s demographic assumptions are used, both polls show Obama well under the 48% level that is presumed to be the lower bound for reelection.

The Gallup poll did not provide a break down of the responses to the presidential approval numbers by party so it was not possible to adjust those results for the differences in demographics.

Rasmussen, who publishes their demographic assumptions has not reported the reversal of position shown by Gallup, who does not report on their demographics. Further this reversal has occurred rather dramatically after Gallup was the target of criticism by the white house. Nothing of note has occurred during the week of so in which Gallup’s numbers flipped.

So at this point it comes down to whose demographics do you believe.

17 Responses to Rasmussen vs Gallup – what’s going on?

  1. ken_phd on April 25, 2012 at 9:19 pm

    Unclefred:

    I guess I don’t see where this false polling will do anything but harm Obama. It certainly won’t stop voters from voting. It may even cause some reduction in effort by the Democrats.

    Is it possible Gallup has one set of results for public consumption and another set they provide to the Dems?

    • tpaine on September 25, 2012 at 10:16 am

      It will give the Democrats something to yell about – another “stolen election” like the one they had at their Convention. Can you imagine. The “anti-God” Party!!

    • Bill on October 2, 2012 at 11:17 am

      Gallup had a poll a month and a half or two months ago and was “called on the carpet” by Axelrod. The DoJ has some lib whistleblower that claims Gallup overcharged for some polling, so the threat of a law suit was made in order to get Gallup “back in line.” That could be why today’s RCP average has Gallup still using “Registered Voters” ( which skews a bit Democratic ) as oppossed to what every other poll has moved to, “Likely Voters” ( which is historically much more accurate ).

  2. Jim,MtnViewCA,USA on April 25, 2012 at 10:02 pm

    I’m with UncleFred on this one.
    And I trust Raz more than Gallup. But hey, I’m just a guy :)

  3. anonymous un-RINO on April 26, 2012 at 4:08 am

    To me, Gallup is a trend pollster. I’d use their stuff for historical purposes, and to gauge against past performances. Their body of work is what makes them useful, not any one poll.

    I’d definitely use them for Obambi’s approval rating, for example, because they have a lot of history with presidential approval ratings.

    However, I wouldn’t bother with them much on discrete political races, between Obambi and Willard for example, because those ARE discrete and not directly leveraged off Gallup’s past work.

    You can’t trust any pollster, really, but I’d trust Gallup whenever they’re dealing with trends and issues. However, they’re not really into politics and the organic interactions therein. They’re a meat and potatoes outfit.

    Rasmussen is well involved with the political free for all, on the other hand. Nobody gets it right in pure political polling, but he’s in the arena credibly. I’d give them a bit more weight in matchup polling in a given race (not that any matchup poll is of much use right now).

  4. setnaffa on April 26, 2012 at 5:39 am

    Do you imagine the polling results are not skewed to influence those who pay the big money to support political campaigns?

  5. Stupid anti-American liberal on April 26, 2012 at 7:15 am

    The Gallup change started IMMEDIATELY after Axelrod attacked their methodology about 8 days ago, or so.

    Go look at it and look at Gallup’s abrupt, inexplicable reversal.

    It’s all there to see.

    • dale on September 8, 2012 at 8:49 pm

      Gee, what else happened in the last 8 days? A disasterous Republican convention, upstages by a hurricane, racial incidents, and a demented actor arguing with an empty chair….and the Democratic Convention, alive with buzz and enthusiasm.

      Here is a report from a conservative organ:

      “Axelrod tweeted in April that a poll showing Republican presidential candidate Mitt Romney with a 48-43 percent lead over Obama was “saddled with some methodological problems,” The Daily Caller reports.

      Clueless Stupid wrote:

      “The Gallup change started IMMEDIATELY after Axelrod attacked their methodology about 8 days ago.”

      Comment: April was not 8 days ago, and “saddled with methodological problems” is hardly an “attack.” You are very confused. The DOJ suit against Gallup is totally unrelated, and goes back to a 2009 Gallup whistleblower filing a lawsuit alleging fraudulent billing of the federal government. Fraud is a crime. However, the Axlerod tweet was 5 months ago and the DOJ has not filed a lawsuit (based on the original complaint from the 2009 whistleblower)..so the allegation of “suspicious timing” falls apart when faced with the real timeline.

      Anything to avoid looking at reality…………..the sinking Republican campaign, the unrelenting popularity of Obama,and the fact that most Americans blame Bush and the current Congress, dominated by Republicans, for the wrong direction the country has taken.

      Romney has been unable to overcome that conviction for the simple reason that his policies are the exact failed policies Bush tried: tax cuts for the rich, deregulation, and more wars. Fake scandals have taken the place of new ideas or solutions for the Republicans. This is the politics of fake outrage and real hatred.

      Read more on Newsmax.com:1.

    • dale on September 8, 2012 at 8:59 pm

      Gee, what else happened in the last 8 days? A disasterous Republican convention, upstages by a hurricane, racial incidents, and a demented actor arguing with an empty chair….and the Democratic Convention, alive with buzz and enthusiasm?

      Here is a report from a conservative organ(newsmax):

      “Axelrod tweeted in April that a poll showing Republican presidential candidate Mitt Romney with a 48-43 percent lead over Obama was “saddled with some methodological problems,” The Daily Caller reports.

      Clueless Stupid wrote:

      “The Gallup change started IMMEDIATELY after Axelrod attacked their methodology about 8 days ago.”

      Comment: April was not 8 days ago, and “saddled with methodological problems” is hardly an “attack.” You are very confused.

      The DOJ suit against Gallup is totally unrelated, and goes back to a 2009 Gallup whistleblower filing a complaint alleging fraudulent billing of the federal government. Fraud is a crime.

      However, the Axlerod tweet was 5 months ago and the DOJ has not yet filed a lawsuit (based on the original complaint from the 2009 whistleblower)..so the allegation of “suspicious timing” falls apart when faced with the real timeline. Solution: fake the dates.

      Anything to avoid looking at reality…………..the sinking Republican campaign, the unrelenting popularity of Obama,and the fact that most Americans blame Bush and the current Congress, dominated by Republicans, for the wrong direction the country has taken.

      Romney has been unable to overcome that conviction for the simple reason that his policies are the exact failed policies Bush tried: tax cuts for the rich, deregulation, and more wars. Fake scandals have taken the place of new ideas or solutions for the Republicans.

      This is the politics of fake outrage and real desperation. As Romney fails to catch up with Obama (especially in the electoral votes, the only ones that count), expect more bogus scandals and fabricated outrage based on more lies hoping to blow up the campaign in a flurry of scandals (like the Republican sex scandals of 2007 and 8, which helped doom the Republcans in that election cycle). But most Americans are tired of fakes, fake conservatives, fake scandals, fake WMDs, fake revolutions, fake news.

      When you find yourself losing, it’s time to cheat, lie, and play dirty. All that matters is winning! And if that fails, you can try to steal the election with voter suppression laws and tampering with the vote count.

      • ChrisGreen on September 13, 2012 at 9:04 am

        “unrelenting popularity of Obama” 5 points behind in one poll and 7 points ahead in another. Gotta love that ‘unrelenting’ popularity. The current economic mess was caused by many of factors, the biggest of which took place in the 1990′s, including government backed insuring of home loans carried by Freddie and Fannie Mae (no matter how bad they were) and the repealing of the Glass Steagal Act. You can blame President Clinton or the Republican congress, I don’t care. However, the idea that the housing bubble was a result only of policy from 2000-2008 is insane. Furthermore, a top marginal tax rate of 38% or 35% would not have prevented the housing bubble either way. I don’t mean this disrespectfully, but I don’t think people were deciding to go with a shady variable interest loan based on a 3% difference in the top marginal tax rate. Wouldn’t you agree?

      • Robin on September 15, 2012 at 6:25 am

        Well said, hanging chads anyone? Talk about trying to steal an election

  6. MikeN on April 26, 2012 at 7:17 am

    Polls are irrelevant this far away from an election. Close to an election, it is pointless to follow them because the actual votes will be happening soon.

    That said, how did you backsolve the Gallup poll? I get the same 6-7 points higher percentage of Democrats, but with a 9% no response, and Independents at 12%, Democrats 3%, and Republicans 7%, that would suggest almost 60% independents in the sample.

    • dale on September 8, 2012 at 9:07 pm

      Polls are critical, both as sources of information and as a tool of persuasion (propaganda).

      That is why all political operatives not only pay close attention to polls but also spend large sums to conduct their own polling.

      Allowing for “house effect (some polls tend to favor Dems, some Republicans)”, polls are samples of the will of the people, subject to change but necessary to understand in order to design rational approaches to changing the results.

      Polls are snapshots of public opinion, often heavily influenced by media propaganda, which is often based on a closes reading of polls. But on average, there is no better way to assess public opinion. And no serious politician or citizen ignores the polls.

      The only poll that counts ultimately is the one on Election Day, but the results on that day are a product of a dialectic involving public opinion, polling, strategies to alter public opinion, and shifts in poll results (or not). Dismissing polling out of hand is ignoring one of the most fascinating and influential institutions in modern politics.

  7. Brian on April 26, 2012 at 7:38 am

    In a prior election (I think it was the last Bush election) Gallup published two tracking polls, one was their traditional methodology and one was their official new methodology at that time. If I remember correctly the old methodology actually was better in the end. Does anyone know if they are still using the old methodology or not?

  8. Trent Telenko on April 26, 2012 at 8:54 am

    A comment from an e-mail acquaintance in the Gallup polling system”

    “I’ve had this “gut” feeling for the past 4 years that Gallup has a democratic BIAS.

    I’m one of their ongoing pollees for the past 8 years but….
    1) they’ve polled me less and less frequently since BO won office and
    2) whenever they do poll me, his numbers go down in the next couple of days.

    Non-scientific observations but the article below seems to bear them out.”

    Make of it what you will.

  9. ChrisGreen on September 13, 2012 at 9:10 am

    Another problem with both presidental polls is that they seem to be polls of all registered voters, not polls of likely voters. I can see how a polling organization can get into trouble making assumptions about likely voters (espeically from a democratic administration). What I’d like to see are 2 polls. One based on all voters. One based likely voters. In fact, 2 or 3 polls based on likely voters should probably be published, each one making slightly different assumptions about the ‘likely’ voter at the time of the election since that metric is really hard to predict and even harder to project into the future.

  10. Phil on September 26, 2012 at 10:44 am

    Rasmussen predicted the mid-term elections in 2010 to within one seat in the Senate and nailed the House. Same in 2008 when they called Obama to win. I tend to trust their polling because they have been right more than the others.

Leave a Reply

Your email address will not be published. Required fields are marked *