Pollster.com

Articles and Analysis

 

Looking for Presidential Exit Polls?

Topics: Exit Polls

Looking for leaked exit poll results from the Obama-McCain presidential race? Sorry to break the news, but until at least 5:00 p.m. today you are out of luck.

Following the 2004 election, when partial and misleading results leaked out at mid day, the network consortium that conducts the exit polls decided to restrict access to a small number of analysts in a "quarantine room" for most of the day. During the primaries this year, and presumably tonight as well, they release their results and vote estimates to producers and reporters at the television networks and other subscriber organizations about about 5:00 p.m. eastern time. While some of that information will no doubt leak after 5:00 p.m, anything you see before that time claiming to be an "exit poll" is probably bogus and certainly not part of the official network exit poll apparatus (Tom Webster, an employee of Edison Research, blogged some details about life inside the quarantine room just before the Super Tuesday primaries).

And while I have your attention, let me offer some advice: Ignore leaked exit polls tonight. I know, I know. How can you ignore them? Everyone wants to know as much as possible about the outcome of this election as soon as possible. But you will do youself a favor if you ignore what leaks out before the polls close, or at least try not to jump to any conclusions about the likely outcome based on what you see. Why? First, the McCain campaign is right: Historically, the leaked exit poll results have "tended to overstate the Democratic vote," and as I reported in March, and the early leaked results during the primaries tended to overstate the Obama vote as well.

Does that information help? Can we apply our own informal adjustment (Obama minus some percentage) and get an precise result? Maybe, but I would not advise it. First, you would still need to consider sampling error, which as always, will be a factor. A margin of a few percentages points means nothing on early leaked results, even if the estimate is unbiased. There is also decent possibility that no such overstatement will occur this time, that it will be smaller than in 2004 or earlier this year (or that what leaks is an already adjusted estimate -- see point #3 below). As Politico reports this morning, new "precautions have been taken to ensure the integrity and accuracy of the exit poll data," including a concerted effort to hire older exit poll interviewers." T

Also, many readers have asked about how exit polls will handle early voting. The short answer is that they have done telephone surveys in 18 states with heavy early voting that will be combined tonight with results from interviews conducted outside polling places. The longer answer is here and here.

And finally, here are a few tips for making sense of the exit poll data you will see tonight (a slightly edited version of tips I posted on the morning of the New Hampshire primary, with a few edits):

1) An exit poll is just a survey. Like other surveys, it is subject to random sampling error and, as those who follow exit polls now understand, occasional problems with non-response bias. In New Hampshire (in 1992) and Arizona (in 1996)* primary election exit polls overstated support for Patrick Buchanan, probably because his more enthusiastic supporters were more willing to be interviewed (and for those tempted to hit he comment button, yes, I know that some believe those past errors suggest massive vote fraud -- I have written about that subject at great length).

2) The networks rarely "call" an election on exit poll results alone. The decision desk analysts require a very high degree of statistical confidence (at least 99.5%) before they will consider calling a winner (the ordinary "margin of error" on pre-election polls typically uses a 95% confidence level). They will also wait for actual results if the exit poll is very different from pre-election poll trends. So a single-digit margin on an exit poll is almost never sufficient to say that a particular candidate will win.

3) Watch out for "The Composite." As they have for the earlier primaries, we expect the web sites of CNN, MSNBC and CBS to post exit poll tabulations shortly after the polls close that will update as the election night wears on (we will post links and commentary here, so we hope you'll plan to check back in later tonight). Those data are weighted to whatever estimate of the outcome the analysts have greatest confidence in at any moment. By the end of the night, the tabulations will be weighted to the official count. Typically, the first waves of exit poll tabulations (including most that leak before the polls close) are weighted to something called the "Composite Estimate," a combination of the exit poll data alone and a "Prior Estimate" that is based largely on pre-election poll results. So if you look to extrapolate from the initial tabulations posted on MSNBC or CNN (as we have done here at Pollster each primary night this year), just keep in mind that in the estimate of each candidate's standing in the initial reports will likely mix exit poll and the pre-election poll estimates (not unlike the kind we report here).

Finally, if you would like more information on how exit polls are conducted, you may want to revisit a Mystery Pollster classic: Exit Polls - What You Should Know.

Now, if you want to know what to make of the exit poll results that the networks will post live on their own sites tonight, you are in the right place. We will be live blogging tonight, and as we did in the primaries, will be reporting on the candidate estimates we can derived from the public tabulations. We also have a special Election Night map that will allow you to do side-by-side comparisons the results with our pre-election poll estimates. So please tune in again later tonight.

Update:  Nate Silver has similar advice, based largely on our own Exit Poll FAQ.

 

Comments
dantheman99:

Thanks for the explaination, Mark.

This was very helpful.

Thank you for your wonderful site.

Dan

____________________

In the 60's and 70's I believe networks used to get actual voting counts from certain precincts around the US and based on those actual counts they could call an election. I would like to see networks get such access and dispense with exit polls.

Please answer this poll on single payer universal health care at http://poll.democratz.org

____________________

Observer:

I haven't seen any analysis of why Obama was overpolled in so many primary exit polls. I did wonder if it was just the nature of the support for each candidate, i.e. Clinton with more older voters, mothers who had to go collect the children etc. Obama with more younger voters prepared to spend the time. And of course early exit polls will underpoll working people who vote after work.

____________________

colorado resident:

Have you heard the Obama reggae song?
This song is truly beautiful! I have been listening to news from around the world and it literally brings tears to my eyes when I think about that one man has the hope of so many people around the world. This is truly a historic moment for the nation as well as the world. You really can't help but to feel a part of something huge, much bigger than yourself when you cast your ONE vote.

Please listen:
http://www.youtube.com/watch?v=lUighoBIx68

____________________

Researchguy66:

@ Observer - As with any intercept research, interviewers are supposed to follow strict guidelines about sampling, usually every "n-th" person during particular time frames. What we notice in consumer research (and I can't imagine is any different in political polling) is that interviewers learn who is more likely to qualify and comply with the request to take a survey and they adapt their own sampling strategy as a result.

It's extremely hard to monitor the behavior of interviewers on a large scale. I've actually hired observers in the past to monitor sampling by interviewers to ascertain if they were following guidelines or not.

That's generally why I won't use intercept for any kind of pre/post or tracking work.

____________________

Vicente Duque:

Thanks for all the hard work. This isn't finished and we hope to follow counting on your studies.

The Early Voting in previous days is very high. Today seems to be good for voting because of fine weather.

There are reports of a big Turnout Today November 4.

All these factor favor Obama.

The Greatest Unknowns :

Missouri for Obama is falling today in INTRADE. Now it has a price of Obama 52.0 and McCain 48.4 ( to win $100 )

North Carolina : Obama 59.8, McCain 49.7

Indiana : Obama 35.3, McCain 66.7

I have total faith in Gamblers and Bettors.

So Indiana is lost for Obama, and Missouri is in Great Danger for Obama.

http://tossUpStates.blogspot.com/

http://milenials.blogspot.com/

Vicente Duque

____________________

Observer:

@Vicente Duque

Gamblers read the polls like everybody else. They don't have any better source of information. A lot of bets were lost when Hillary Clinton won the New Hampshire Primary.

____________________

Vicente Duque:

Dear Observer :

I am a "Casino Junky" ( just kidding ), I have the strongest faith in Gamblers. They are inside the campaigns, inside Government, inside counting votes. I keep a constant watch on them.

Breaking News !

Obama is increasing bet value in these three states.

North Carolina (15EVs) .. Obama 64.4 McCain 41.8

Missouri (11EVs) ......... Obama 58.8, McCain 38.5

Indiana (11EVs) ........... Obama 43.5, McCain 57.3

It is amazing how Obama is increasing price in these three states.

These bettors are great cheaters. They have "inside information". They see people smoking, drinking, mad or weeping and draw conclusions.

I am constantly updating the prices of bets at INTRADE, here :

http://tossUpStates.blogspot.com/

http://milenials.blogspot.com/

Vicente Duque

____________________



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