JONATHAN CHAIT MAY 25, 2010
-
Read Later
READ LATERAvailable only to subscribers. SUBSCRIBE TODAY
-
Listen
ARTICLE AUDIO
- Font Size
Daniel Foster at National Review trumpets a decline in President Obama's approval ratings:
Obama Approval Index Hits New Low
Rasmussen has Obama's approval index — strong disapproval subtracted from strong approval — at -20, its lowest point to date. Obama's overall approval/disapproval split stands 42/56, dragged down by poor grades on the economy and the handling of the BP oil spill.
Are Obama's approval ratings falling? Well, yes, according to Rasmussen. No, according to everybody else:
Rasmussen polling occupies an odd place in the political culture. In the conservative world, it is the gold standard. If you go to a conservative set on basically any random day, you'll see somebody touting a Rasmussen poll. Here is John McCormack at the Weekly Standard touting a poll showing huge support to repeal the Affordable Care Act. Here is Peter Wehner at Commentary doing the same. Rasmussen frequently asks unusual polling questions that produce results almost certainly calculated to demonstrate public support for the conservative position. (Here's one example of a loaded Rasmussen question.) Rasmussen has become a right-wing celebrity. He's the author of a conservative book. This fall he is a featured guest on National Review's cruise, along with other conservative luminaries.
Part of Rasmussen's celebrity status derives from the fact that even his polls on commonly-asked questions skew strongly toward the conservative position. Here, for instance, is Nate Silver's depiction of Rasmussen's measure of party identification:
Rasmussen's role in the public debate is problematic for several reasons. It's not altogether clear what causes him to consistently project results so much at odds with those of the rest of the polling community. But if there is something problematic about his methods, he has little incentive to correct it, because Rasmussen's business model increasingly relies upon maintaining the loyalty of staunch Republicans.
Now, to be perfectly clear about this, it's possible that Rasmussen is right and everybody else is wrong. The safest approach to using polling data is to include all results. The trouble is that Rasmussen can have such large outliers, and it polls so often, that the very inclusion of Rasmussen changes the results. The graph near the top of this item, showing level public approval for Obama, would show a steep dip if it included Rasmussen's findings, the latest of which has Obama sporting a disastrous 42/56 approval rating.
But the more problematic dynamic is Rasmussen's symbiotic relationship with the conservative base. The habitual practice by conservative pundits of quoting only Rasmussen polling reinforces conservatives' overweening certainty that they embody public opinion. It's an important component of right-wing epistemic closure, the Republican base having its own pollster who always tells them what they want to hear. In theory, there ought to be a corrective dynamic. If Rasmussen is wrong about the 2010 elections -- and, again, you can't be certain he will be -- in theory, this would cause Republicans to question their reliance upon his unusual findings. But it's entirely possible that Republicans would simply question the validity of the results themselves. It's massive voter fraud! Obama dirty tricks! Having heard on a daily basis that the American public had rejected the Democrats wholeheartedly, disbelieving the validity of the election results would create less cognitive dissonance.
Of course, one solution would be for the conservative pundits who relentlessly cite Rasmussen's findings to inform their readers that those findings, while not necessarily wrong, represent an outlier among pollsters. But I wouldn't hold my breath waiting for that to happen.
RELATED: Click here to read Chait's follow-up post about Rasmussen.
19 comments
It is a fool's errand to trust Rasmussen's numbers. It is frequently an outlier, often out of sync with the average of whatever is being polled by the quality pollsters. Right-wing ideologues love Rasmussen because it spins the numbers their way. The smart ones know this and therefore are disingenuous when they cite Rasmussen to bolster their arguments. The fatuous ones are gullible and aren't aware of this elementary fact.
- liberal reformer
May 25, 2010 at 3:13pm
I appreciate being able to comment on this thread. The Sestak Pseudo-Scandal post has no provision for comments. This is the third time in two days that this has happened. What gives?
- liberal reformer
May 25, 2010 at 3:19pm
How does Rasmussen compare to other pollsters on predicting the outcomes of past elections?
- Simon Greenwood
May 25, 2010 at 3:22pm
Rasmussen's last pre-election poll in 2008 over-reported support for McCain/Palin, but Rasmussen nonetheless came closest to the actual outcome of the election. See http://bit.ly/dhi3cm This year, Rasmussen has not polled in any of the interesting elections close to election day, so it's not possible to test Rasmussen polling results against actual election results. Convenient, that.
- rhubarbs
May 25, 2010 at 3:32pm
- "In the conservative world, it is the gold standard.". As in a myth, folklore, fable or fallacy? You said it all in a single sentence.
- michaelg
May 25, 2010 at 3:52pm
While I usually take Kos with somewhat of a grain of salt, he posted an interesting theory on Rasmussen. http://www.dailykos.com/storyonly/2010/5/19/867829/-Rasmussens-(dishonest)-game Ras is all about setting a narrative. When compared with results of major elections, Rasmussen generally does pretty well on polling. However, very little of Rasmussen's released polls deal with elections, rather most of them survey quirky or out of the mainstream issues/races. Since a lot of the polling topics are odd or unique they can't be compared to other polling results. Since Ras is the only outfit polling these questions or obscure races they flood the zone with their numbers and hence can set the narrative or at least provide the groundwork for other conservative pundits. Another way Ras sets a narrative is that they poll a race for a period prior to the election and then pull out weeks before the actual election happens. That way they can't be held accountable if the poll is way off. This gives them the ability to release a mountain of polls on a given topic/race and set a general narrative, but never have to design that poll for accuracy. It is well known that the likely voter screen that Ras uses selects heavily for Republicans. On many of these questions or races Rasmussen does not release crosstabs, so there is not a way to actually scrutinize the poll for legitimacy. On major election polling, Ras does release crosstabs, and their results are much closer to other major polls. By rigging their polling system to produce accurate results on major questions which can be tested against actual elections they build the "cred" that allows them to release bogus polls that are purely narrative setting and that can never be empirically proven. Nate Silver has written on this as well. http://www.fivethirtyeight.com/2010/05/paulapolooza.html
- skottyj
May 25, 2010 at 4:07pm
lib ref. Chait has a polling problem, too--- in his "Halter is toast" post, to which no reply is possible. The 538 poll analyses are WAY out of date. Does Chait not know that? Do you really think that Toomey has a 71% chance of winning PA against Sestak -- or the Repub a 78% chance of winning FL -- or the Chicken Lady or the Tea Bagger a 77% chance to beat Reid--or Rand Paul an 83% chance to win KY. I'd sure take that combination against 538 calculations not updated for almost a month--- and give Halter more like a 30-40% chance against his opponent. Them's not the best odds, but they ain't hopeless.
- drofnats1
May 25, 2010 at 4:09pm
"The safest approach to using polling data is to include all results." Ok, my Richard Serlin poll shows 12,000% of Americans approve of President Obama. Please average that in with the other results. Allright, so statistics says you weight results, or data, by their reliability (to the extent it's practical), and try to correct for their biases. Right now it's looking like Rasmussen results should be weighted very lightly and/or corrected for their large bias.
- RHSerlin
May 25, 2010 at 6:27pm
Jonathan Chait doesn't do polling, Rasmussen does.
- liberal reformer
May 25, 2010 at 6:43pm
Chait reports on polling.. Rasmussen and 538. In the first instance with many caveats, in the second-- swallowed whole.
- drofnats1
May 25, 2010 at 7:04pm
538, as well as many other pollsters, are more reliable than Rasmussen. You start suspecting either the methodology or the ideology (or both) of a pollster whose polls are so often outliers. Didn't you know?
- liberal reformer
May 25, 2010 at 9:03pm
Or it could be that the Republicans will continue to peddle Rasmussen numbers because they (the numbers) allow them to reinforce their viewpoints - no one will remember how Rasmussen did in predicting a particular election or issue other than Nate and people who read him (and TNR). Isn't Jonathan being a little naive, assuming that Republican politicians actually want their polls publicly validated? In all likelihood, their number crunchers are giving them the correct forecast based on all polls - it's just that they choose to trumpet the bent ones. That way, when reality doesn't match their publicly claimed expectations, they can claim voter fraud, etc. It's a gift that keeps on giving ...
- NR409654
May 25, 2010 at 9:44pm
Chait just doesn’t get it. How stupid do you have to be to qualify for liberal blogger? Isn't there a floor? Rasmussen is polling Likely Voters, almost all the other outfits are polling adults or Registered Voters. It doesn't stretch the imagination to understand that voters who disapprove of Obama -- productive private sector types who will be paying for his stupidity -- are more likely to vote than liberals. This is probably why Rasmussen does very well in predicting elections -- likely voters are better predictors than registered voters or adults.
- mr_rationale
May 26, 2010 at 9:17am
Almost all polling operations survey "likely voters." Since there's no way to know who is actually going to vote until after the election, estimating whether any individual is a "likely voter" is educated guesswork. In order to survey "likely voters," Rasmussen must put his finger on the scales and tilt them one way or another. Rasmussen makes assumptions about voter behavior based on factors like race, education, and party, and tilts his polls very strongly one way to arrive at a "likely voter" model. Other pollsters make different assumptions, or tilt their surveys less strongly, to arrive at their own "likely voter" models. Also, most of Rasmussen's polling is on public opinion generally, and on most of those polls Rasmussen does not impose a "likely voter" screen. To suggest that using "likely voters" makes Rasmussen different from other pollsters, or makes Rasmussen's data more reliable, is simply to declare one's complete ignorance of how opinion polling, as well as basic statistics, works.
- rhubarbs
May 26, 2010 at 9:40am
The real problem is when Republican's get into office and ignore reality when it comes to public policy. If they are deluding themselves about their electoral chances then that's fine with me. Let the complacency set in.
- poldpf
May 26, 2010 at 10:27am
To add to Rhubarbs' point, most reputable pollsters (of whatever idealogical stripe) have eschewed giving a lot of credence to polls of "likely voters" more than three months before the actual election in question. There is too much of a chance that the people who say they are "likely voters" six or twelve months out will change their mind (especially if they are exposed to sustained TV or signage advertising from the other guy closer to the election), move, get sick, die or be otherwise indisposed to vote in the real election. There are some arguments out there that a persistent focus on the "likely voter" makes more sense in today's Internet and cable-news driven political environment, but I recall that the evidence from the last three election cycles is that they don't do a very good job of predicting real outcomes.
- wildboy
May 26, 2010 at 4:15pm
Polling likely voters makes sense when polling on an impending election, otherwise it is just a smokescreen to cook the books. On general issues of public policy or presidential approval, all citizens are equal and should be considered equally, whether they vote or not. Voter participation also fluctuates between local and national and Presidential elections. Public opinion polling is done to get the opinion of the public, not the opinion of some select group you are guessing might vote in 2012.
- nayyer_ali
May 26, 2010 at 5:55pm
rhubarbs -- you are like a clone of Chait -- stupid and unaware Fact: Almost all polling operation DO NOT survey likely voters. Of the following list of polling firms surveying O approval only Rasmussen uses LV (Likely Voters). The remainder use A (Adults) or RV (Registered Voters) as their sample Gallup 5/24 - 5/26 1547 A 46 47 -1 Rasmussen Reports 5/24 - 5/26 1500 LV 45 53 -8 CBS News 5/20 - 5/24 1054 A 47 43 +4 Quinnipiac 5/19 - 5/24 1914 RV 48 43 +5 CNN/Opinion Research 5/21 - 5/23 1023 A 51 46 +5 NBC News 5/20 - 5/23 700 A 48 45 +3 FOX News 5/18 - 5/19 900 RV 45 46 -1 Democracy Corps (D) 5/15 - 5/18 1000 RV 47 47 Tie Associated Press/GfK 5/7 - 5/11 1002 A 49 50 -1 NBC News/Wall St. Jrnl 5/6 - 5/10 1000 A 50 44 +6 Ipsos/McClatchy 5/6 - 5/9 1016 A 52 43 +9 Pew Research 5/6 - 5/9 994 A 47 42 +5 1 out of 12 polling organization using likely voters. 11 out of 12 not using likely voters You are utterly wrong. If you want to refute, please use your brain (if you have one)
- mr_rationale
May 27, 2010 at 3:14pm
With very little background in polling, it just seems odd to me that Rasmussen is consistently a mirror image of the other polls. Most other polls I have seen, for the most part show slightly positive ratings for Obama, which is roughly high 40's to low 50's approval vs. low to mid 40's disapproval (although this has gone down somewhat recently with the oil spill). Rasmussen shows just the opposite - all the time. So all those other polls, done by different organizations are wrong, but Rasmussen is right? It would be interesting to see the daily tracking numbers without Rasmussen considered.
- RobertW
May 28, 2010 at 12:12am