Editore"s Note
Tilting at Windmills

Email Newsletter icon, E-mail Newsletter icon, Email List icon, E-mail List icon Sign up for Free News & Updates

January 9, 2008
By: Kevin Drum

EXIT POLLS....This is kind of dumb and dorky, but it's worth sounding a little bit of a warning about the exit poll results from New Hampshire. (Or any other exit poll results, for that matter.) Here's the warning: a lot of the subgroups are pretty small, which means the margin of error is huge. Take a look at the age breakdown in the Democratic primary, for example:

Now, it's possible that Barack Obama really did win the 18-24 and 30-39 year-olds but that Hillary Clinton inexplicably appealed to the middle group of 25-29 year-olds. But it's not likely. What is likely is that it's just a statistical artifact: the 25-29 segment is small and the margin of error among that group is a whopping 8%. There's a pretty good chance that Obama actually won those folks, just like he won the surrounding groups, but that random chance produced the opposite result.

Anyway, this is your dumb and dorky stats post for the day. Just a warning not to take the most detailed parts of the exit polls too seriously.

Kevin Drum 12:51 PM Permalink | Trackbacks | Comments (18)

Bookmark and Share
 
Comments

Actually, looking at that I would also lean towards thinking that Obama's 30-39 year old number is overstated, so that Obama may have won both categories but by a much closer margin than indicated.

Posted by: Jimm on January 9, 2008 at 1:01 PM | PERMALINK

Also, I heard on the news last night that Hillary did spectacularly with single women - all women really but best with single women - so maybe that explains the 25-29 outlier as well (since you probably have the most single women in the 18-24 and 25-29 brackets, and perhaps the youngest voters are the most gung-ho about Obama).

Posted by: Jimm on January 9, 2008 at 1:03 PM | PERMALINK

Kevin-

I agree. It's safe to say that Obama did well among young voters, independents, and higher educated/higher income voters and Hillary did well among women, older voters, and lower educated/lower income voters.

As an Obama supporter this scares me. If Hillary continues to get the women vote, the 50-and over vote, and the working class vote she will win the nomination. That's a powerful bloc on the Democratic side. Obama's coalition, energetic as it is, is simply not big enough unless he can cut into the women vote and convince older, risk-averse voters that he can be a good leader.

Posted by: David68 on January 9, 2008 at 1:05 PM | PERMALINK

"There's a pretty good chance that Obama actually won those folks, just like he won the surrounding groups, but that random chance produced the opposite result."

Actually its reasonably plausible that the exit polls are accurate.

It resonable to assume that the 18-24 segement was heavily represented by people in college. This group may have demographic difference more stark than age would suggest.

In anycase, dont over think it too much. Obama tended to do well with younger voters (probably especially college age) and Clinton tended to get more older voters.

Posted by: Catch22 on January 9, 2008 at 1:11 PM | PERMALINK

Obama beat Clinton by 2 votes in the combined category of voters 18-29.

Posted by: Yancey Ward on January 9, 2008 at 1:32 PM | PERMALINK

I am a litte rusty on my stats but winning 37-35 is pretty close to a dead heat with a margin or error that big.

It is obvious from the trend that Clinton did better with age and Obama did worse with age.

You basically have two choices, either Clinton 'probably' won the 25-29 vote and there is something unique about that age group (like they were a certain age during an important time in the Clinton presidency) or
Clinton 'probably' didn't win and you have a statistical fluke that occurs about 5% to 10% of the time.

I vote for choice #2

Posted by: neil wilson on January 9, 2008 at 1:38 PM | PERMALINK

All I can say is that I'm a 25-29 single woman who would never ever vote for perpetuation of the Bush-Clinton dynasty, headed by a spineless DLCer who spent the last seven years enabling the Bushies and not much else.

Just why Kevin Drum thinks she'll be a good president? I guess that's the sort of blind ignorance of reality one would expect from someone who is still shocked when the conservatives do something evil or stupid.

Posted by: anon on January 9, 2008 at 1:43 PM | PERMALINK

What share of the NH voting population do Hillary's older constituents represent?

Is Hillary benefiting from the growing numbers of aging baby boomers?

Posted by: pj_in_jesusland on January 9, 2008 at 1:50 PM | PERMALINK

(Or any other exit poll results, for that matter.)

So, Kevin: Have you also drunk the Kool-Aid regarding the overall reliability of exit polls, historically established at better than +/- 1.0%?

There was NOTHING unusual about Kerry's supposed 2004 "loss" being so widely divergent wirh the recognized accuracy of those exit polls??

Well, hell. I've got a bridge I can sell ya!
.

Posted by: Poilu on January 9, 2008 at 2:02 PM | PERMALINK

My guess is that the biggest artifact is the 18-24. I've worked extensively with that group for about the last 7 years and have been immersed (undergrad and grad school) for approaching 10 years. I don't doubt Obama did well in that group, but 3:1 seems way too high. Everyone thought Obama would come back once the college precincts came in, but he didn't. That suggest to me that his prowess in those precincts was overstated. Over the last several elections, 18-24 wasn't as dramatically different from 18-29 as these exit polls suggest.

The Obama supporters may be loud and boisterous and extremely committed (as were Dean's), but just as Dean's perceived support among young people didn't really match reality I don't think Obama's does either. Committed activists are most likely way in favor of Obama, but not the overall population. Obama's young supporters were probably more willing to answer the pollsters than nonsupporters.

Posted by: gq on January 9, 2008 at 2:05 PM | PERMALINK

I'd call it the Chelsea Effect. Hillary's daughter is out campaigning for her, and her daughter is smack dab in the middle of that 25-29 age group.

Posted by: MDS on January 9, 2008 at 2:07 PM | PERMALINK

From the profile of New Hampshire primary voters, one factor that will have greater importance in the future is the category of VoterID: Clinton won regular party voters 45 to 34%. Many upcoming states do not allow cross-over voting.

One of Obama's problems is that he is new to the national scene. Voters, especially older voters, are more comfortable with people they're familiar with or whose stories are better known. An extended primary season allows new candidates to get their story out there; this compressed election doesn't.


Posted by: Mike on January 9, 2008 at 2:33 PM | PERMALINK

Poor Gravel.

Posted by: Joshua on January 9, 2008 at 2:35 PM | PERMALINK

her daughter is smack dab in the middle of that 25-29 age group. Yeah, and a fox to boot! Just like her Mom.

Posted by: Tripp on January 9, 2008 at 2:45 PM | PERMALINK

Hillary's daughter is out campaigning for her

Is she? I thought she was staying out of the campaign.

Yeah, and a fox to boot!

It's gotta be killing Rush Limbaugh that Chelsea turned out so pretty after his years of cruelly ragging on her looks, while he, already visually disgusting during the Clinton era, now resembles an army-size bag of library paste.

Or it would if he had any self-awareness at all.

Posted by: shortstop on January 9, 2008 at 4:05 PM | PERMALINK

      The media was so concerned with what went wrong with the polling in NH, as if the credibility of the media was on the line.  Nothing went wrong.  The statistics didn't lie either; it is the nature of factual questions that fail to engage one's inner passion.  The American phenomenon is how people poll with their brains, but vote with their hearts.  This race is anything goes because the American people believe anyone can be President.  That is especially true for the first open election in 80 years.  (Personally, I am still holding out for someone to come on the scene who has the answers.)  After we get to know Obama better than he knows himself, and Rahm Emanuel's words come back to haunt Hillary, then Edwards will be picked out of spite.  As for the GOP, Huckabee is tripping over Woodrow Wilson while preaching on the way how we must endorse the League of Nations, McCain is tripping over Iraq and Immigration, and when Giuliani tripped no one noticed.  When us Dems come up short for a lack of a unified plan, my guess is Edwards will join forces with a yet unnamed business-strategist at nomination time to defeat turnaround specialist Romney.  However, without the full backing of all the major contenders as co-presidents, Edwards will look silly scolding Americans for the poverty-stricken among us when the Kondratieff Economic Winter comes to demoralize its homeowners and SUV owners.  Romney may flop like a fish, but he is the farthest upstream come November.  Tell Rahm to start making some phone calls and invite some bears to go fishing.

Posted by: Stephen L. Rush on January 11, 2008 at 2:42 AM | PERMALINK

      The media was so concerned with what went wrong with the polling in NH, as if the credibility of the media was on the line.  Nothing went wrong.  The statistics didn't lie either; it is the nature of factual questions that fail to engage one's inner passion.  The American phenomenon is how people poll with their brains, but vote with their hearts.  This race is anything goes because the American people believe anyone can be President.  That is especially true for the first open election in 80 years.  (Personally, I am still holding out for someone to come on the scene who has the answers.)  After we get to know Obama better than he knows himself, and Rahm Emanuel's words come back to haunt Hillary, then Edwards will be picked out of spite.  As for the GOP, Huckabee is tripping over Woodrow Wilson while preaching on the way how we must endorse the League of Nations, McCain is tripping over Iraq and Immigration, and when Giuliani tripped no one noticed.  When us Dems come up short for a lack of a unified plan, my guess is Edwards will join forces with a yet unnamed business-strategist at nomination time to defeat turnaround specialist Romney.  However, without the full backing of all the major contenders as co-presidents, Edwards will look silly scolding Americans for the poverty-stricken among us when the Kondratieff Economic Winter comes to demoralize its homeowners and SUV owners.  Romney may flop like a fish, but he is the farthest upstream come November.  Tell Rahm to start making some phone calls and invite some bears to go fishing.

Posted by: Stephen L. Rush on January 11, 2008 at 2:48 AM | PERMALINK

nline backgammon resource and magazine with hundreds of backgammon articles and matches.

online backgammon - http://www.nacr.net/ Posted by: online backgammon on February 10, 2008 at 1:59 PM | PERMALINK




 

 

Read Jonathan Rowe remembrance and articles
Email Newsletter icon, E-mail Newsletter icon, Email List icon, E-mail List icon Sign up for Free News & Updates

Advertise in WM



buy from Amazon and
support the Monthly