Didn't Clinton and Obama tie in New Hampshire?

January 10, 2008 9:25 AM

If the point of a Presidential Primary is to determine how many delegates will vote for a given person in the next phase of picking a nominee, and both Hillary Clinton and Barack Obama got 9 pledged delegates in New Hampshire, didn't they tie?

I'm not necessarily for Clinton or Obama, and I don't want to talk about the merits of any candidate. I'm just curious about the mechanics of this one thing, and why, even though they both got 9 delegates, Hillary Clinton is the winner of New Hampshire.

posted by bryanjbusch to law & government (23 comments total)

explained in slate.
posted by Hat Maui at 9:31 AM on January 10, 2008

Because they didn't tie--Hillary got more votes.

Perception matters more than delegates earned at this early stage in the nominating process, because of the momentum having a win gives a candidate. I.e., Clinton is now seen as more electable and viable than she was before New Hampshire, and that makes voters in future states more likely to support her.
posted by dyslexictraveler at 9:33 AM on January 10, 2008

The secondary purpose of a primary to is judge how a candidate might do in the general election. In that sense, Clinton won because she got a higher percentage of the vote, and it was just following a loss in Iowa where she was expected to do better.

What if every primary ended like that? If the final delegate count was tied, but one candidate won most of the popular votes in all of them, that candidate would be the obvious choice at the convention.
posted by InfidelZombie at 9:34 AM on January 10, 2008

Hillary won the popular vote.
We all learned in 2000 that the popular vote doesn't really matter.

I'm a Kucinich supporter myself and really don't like the way the media is handling this election.
posted by Dillenger69 at 9:34 AM on January 10, 2008

It's much more exicting for the media to say any one person "won" a state (as if it's the election already and they can "win" that state's portion of the electoral college), when in fact, there really isn't a lot of winning when, as you point out, they both have nine delegates. It only matters who "wins' a state when the race isn't this close. It's much more accurate to talk about Clinton/Obama in terms of how they are doing delegate wise, which unless I'm completely mistaken, is fairly close. IE, this is a good, competetitive race in that regard. If you like neither Hillary nor Obama, then you may disagree.
posted by Medieval Maven at 9:38 AM on January 10, 2008

I'll play along with the "secondary purpose" angle, but I also sympathize with Dillenger69 when he/she says that we learned in 2000 that the popular vote doesn't actually mean much.
posted by bryanjbusch at 9:38 AM on January 10, 2008

we learned in 2000 that the popular vote doesn't actually mean much.

Uh, it does on a state-by-state basis.
posted by grouse at 9:43 AM on January 10, 2008

they tied delegatewise but not popular votewise, and america was watching to see if she could get up off the mat after obama clobbered her in iowa. it's still a fight!
posted by bruce at 9:44 AM on January 10, 2008

That's true. It's also interesting that Republicans and Democrats handle delegates differently. It's a good thing we have journalists to boil it down to "these two people won."
posted by bryanjbusch at 9:45 AM on January 10, 2008

[please do not turn this into a referendum on the 2000 election or your own personal feelings on who should be getting votes, thanks]
posted by jessamyn at 9:52 AM on January 10, 2008

Yes, it's a tie.

No, but doesn't make for good headlines.
posted by Hugh2d2 at 9:55 AM on January 10, 2008

I was doing my best, jessamyn. I think the thread's complete now, anyway.
posted by bryanjbusch at 9:55 AM on January 10, 2008

Also, don't forget that the Democratic party has the concept of Superdelegates that aren't controlled by the outcome of the individual state races. Winning a majority of votes in states is one of many ways to influence how these superdelegates vote at the convention.
posted by mmascolino at 10:31 AM on January 10, 2008

See, this brings up something I was sort of wondering about, but that I don't have the mathematical knowledge to determine one way or another: How do the concepts of statistical significance and/or margin of error play into the results when the percentages are so close? Or are those concepts even relevant when it comes to the actual primary election voting process, given that

1. One's choices are discrete (you can only vote for Candidate 1, Candidate 2 or Candidate 3, and only on one ballot, Democrat or Republican)

2. The data collection is so standardized (all data comes in via ballot)

and 3. The number of delegates is only allowed to be a whole number

?

I guess my nutshell question, coming from a psychology/statistical background, would be:

Was the difference between percentages voting for Obama and Clinton in the New Hampshire primary statistically significant? And if so, how large was the effect size?
posted by limeonaire at 10:52 AM on January 10, 2008

I don't think that statistical margin of error comes into play because the primary isn't a poll. That is, it isn't an attempt to estimate something the way a poll is.
posted by XMLicious at 11:04 AM on January 10, 2008

In other words, it isn't taking a sample of a population, it IS the population, in its entirety.
posted by XMLicious at 11:06 AM on January 10, 2008

limeonaire-- it's not a survey based on a random sample, it's an actual count of all the ballots submitted. If I have 100 people in a room and ask them all their opinion on something, that's different than asking 30 of them and determining via statistics what the other 70 likely think. Margin of error and significance don't apply in this case.
posted by InfidelZombie at 11:09 AM on January 10, 2008

Statistical significance and margin of error only come into play when you're sampling a population, that is, polling a certain subset of the population and attempting to use it to model the entire population. It doesn't really apply to the actual election.

An analogy: suppose you have a bag of a million marbles, some white and some black. You want to know what percentage of the marbles are white. But suppose you don't want to put in the effort to count a million marbles--you decide you're just going to take a thousand marbles at random out of the bag and take that as a representative of the whole set of a million marbles. Let's say 328 of the marbles are white. You conclude that 32.8% of the marbles in the bag are white, but you also say you have a 3% margin of error. This means that usually (to be precise, 95% of the time), when you take a sample of a thousand out of a population of a million, the percentage that you calculate based on your sample is within 3% of the true value--what you would get if you did count all million marbles. (Aside: one of the counter-intuitive results of statistics is that the margin of error depends almost entirely on the sample size but very little on the population size, except in cases where the total population is only slightly larger than the sample. A sample of one thousand has a 3% margin of error, regardless of whether it's one thousand out of ten thousand, one thousand out of a million, or one thousand out of a billion.)

But if you actually did go through and count all million marbles, there would be zero margin of error, because you've counted all there are. Important: this is not the same as saying your count is exactly correct; you can make mistakes after all, and maybe in the tedium of counting all of those you miscounted some white as black, or vice versa. But "margin of error," as a statistical term, does not cover that sort of error. Saying there is zero margin of error does not mean your count is exactly right: it only means there's no error due to the fact that you're trying to take a sample as representative of the whole. Likewise, there is no margin of error in the election itself, because the election isn't a poll, the election is the actual count of everyone. Which is not to say that there are no errors in the election results at all, only that the concept of "margin of error" cannot be used to measure the type of error which might occur in an election.
posted by DevilsAdvocate at 11:20 AM on January 10, 2008 [1 favorite]

Ahh, right. I knew there was a flaw in my logic somewhere, I just couldn't figure out what I was missing. Duh—it's not sampling data. Thank you, XMLicious, InfidelZombie and DevilsAdvocate.

Sorry, I'm the girl who ended up taking psych stats twice in the course of getting her (largely work-irrelevant) psychology degree... I just needed a quick correction.
posted by limeonaire at 11:28 AM on January 10, 2008

In the Democratic party, about 20% of the delegates are party wheelhorses who do not formally represent any state, and who decide for themselves who to vote for. They're referred to as "super delegates".

In some campaign years that doesn't matter. Some candidate will get enough delegates in caucus and through primaries to cinch the nomination. But if no candidate does so, then these uncommitted delegates will ultimately decide who gets the nomination, and among other things they'll base their decision on issues of electability in the general election. That will include examination of how actual voting landed in various states and regions.

It's been decades since the last time the nomination was in doubt at the convention, but it's possible if the race is tight.
posted by Steven C. Den Beste at 11:31 AM on January 10, 2008

This news article states, "Clinton leads with 187 delegates, including separately chosen party and elected officials known as superdelegates. She is followed by Obama with 89 delegates and Edwards with 50." Therefore I get impression that the positions of the superdelegates are already established?
posted by yeti at 11:51 AM on January 10, 2008

List of superdelegates that have officially announced who they plan to nominate.
posted by Guy Smiley at 12:09 PM on January 10, 2008

Super delegates are not formally dedicated, but there are some who have already announced their allegiance. However, delegates from many states are required by state law to vote for their primary winners. No such compulsion exists for super delegates, and those currently counted for Hillary could change their minds later.
posted by Steven C. Den Beste at 12:11 PM on January 10, 2008