Bell Curve The Law Talking Guy Raised by Republicans U.S. West
Well, he's kind of had it in for me ever since I accidentally ran over his dog. Actually, replace "accidentally" with "repeatedly," and replace "dog" with "son."

Thursday, February 07, 2008

Estimate: Hillary +100 in Pledged Delegates

The NY Times is conservative but up-to-date in reporting delegate totals. For the states they tally, they are more complete than CNN, but they omit several states. NY Times explains, "The New York Times counts only delegates that have been officially selected and are bound by their preferences." (And of course they count nothing for Michigan or Florida for the Democrats.)

As of 7:45 PST 2/7/08, NY Times shows 892 for Hillary and 716 for Obama without any superdelegates. Yes, those are pledged delegates only. But wait! They have not yet tallied results from seven states, all of which favored Obama in delegates. (They are also waiting for final results from American Samoa.) But we can add estimates for pledged delegates for these states from CNN. I show them explicitly here to avoid confusion.

STATE/TERR.: Obama, Clinton
Am. Samoa: 1, 2 (Wikipedia)
Alaska: 9, 4
Colorado: 37, 18 (CNN says 13, 6 but omits 36. I assume 2-1 split for Obama)
Idaho: 15, 3
Iowa: 16, 15
Minnesota: 48, 24
Nevada: 13, 12
North Dakota: 8, 5

Adding in these results, the more complete delegate estimate is therefore: Clinton, 975; Obama, 876. Then you need to add in the superdelegates on top of that. NY Times says 99, 204; CNN says 106, 193. I'll use the more even CNN estimate.

DELEGATE TYPE: Obama, Clinton
Pledged: 876, 975
Super: 106, 193
TOTAL: 982, 1168
NET: +186 for Hillary

In other words, Super Tuesday was not a tie but really a modest victory for Hillary: she earned about +100 net in pledged delegates, to add to her existing +100 edge in superdelegates.


The Law Talking Guy said...

FYI: Real Clear Politics has different numbers. 1060 Clinton (includes 211 Superdelegate) and 981 Obama (includes 128 superdelegates). Subtracting the superdelegats, that's 849 Clinton, 853 Obama.

I think what's happening is that the delegate allocation process and reporting is not uniform, particularly in caucus states. It may be a few days before the numbers shake out.

Dr. Strangelove said...

Given the NY times' procedure, I trust their numbers.

The Law Talking Guy said...

It's not about who you trust, it's about how final the numbers are. I don't think we have final numbers yet. May take a couple days.

The Law Talking Guy said...

CNN has 783 clinton, 760 obama.

Washington Post says: 845 Clinton, 765 Obama, including superdelegates.

MSNBC says: 861 Obama, 855 Clinton.

Dr. Strangelove said...

Unless NY Times lied about how they compile their figures (see above), or they made an error in data entry, their numbers are final. (My handful of additions via CNN were only estimates, but are probably close.)

Dr. Strangelove said...

UPDATE: NY Times pledged delegate count now has 904 for Clinton and 724 for Obama. New Mexico is now in. For some reason, I failed to add in any estimate for this above. There are no new updates from CNN.

My augmented estimates for pledged delegates ONLY (now including the formerly missing New Mexico results) are thus:
Clinton, 989; Obama, 888.

Dr. Strangelove said...

If (hypothetically) the other news organizations catch up to the NY Times figures, I will have to wonder why they chose to bury the story that Hillary came out a bit in front. I mean, it's not like they are slow to jump to conclusions about any other aspect of the race.

If that proves to be the case, one would have to wonder whether the demographics of journalists--wealthy, white males--had anything to do with it. Those tend to be Obama supporters. (Vast left-wing conspiracy, anyone? Just kidding.)

The Law Talking Guy said...

NY Times writes:

"Many news organizations include delegate projections in their counts that are based on nonbinding votes for candidate preference, such as the Iowa caucuses. The New York Times counts only delegates that have been officially selected and are bound by their preferences."

So, Iowa and Nevada are not in these numbers. Also, some states have a multi-stage process, so the delegates are counted differently. For example, Kansas actually has 41 delegates, not the 21 that the NYTimes relies on. 21 were chosen on Super Tuesday. The next 20 are chosen by a complicated process, but largely by the groups already selected at the first-stage caucuses to go to the state convention (and they're overwhelmingly Obama supporters). So that's why CNN, for example, counts these delegates as already-selected for Obama, while NYTimes doesn't. These complicated rules (see explain why delegate counts differ. They also explain that the NYTimes total is a bit deceptive, since in caucus states, the delegates may not technically be selected, but their composition is often a foregone conclusion that both campaigns know about.

The Law Talking Guy said...

There are also rules about superdelegates that are worth noting. Some are pledged and some are unpledged, and the pledged ones are sometimes counted and others not.

Dr. Strangelove said...

Well, we'll see how it irons out. For some reason, NY Times has 80-100 more pledged delegates for Hillary than anyone else, and their count is the most conservative. Still can't wrap my head around that one.

I accounted for the presumed pledged delegates from the seven or so races omitted by the NY Times, but I did consider the additional delegates that are presumed to follow from in the multi-stage caucus process. Good point there.

My only puzzle is that adding in those extra second-stage delegates could only increase the vote totals, whereas the others have tabulated fewer total pledged delegates than the NY Times.

Maybe we'll see some convergence soon. Until then, I'm sticking with the more conservative estimates.

The Law Talking Guy said...

I'm not sure why the estimates that favor HRC are "more conservative." Or why that's a virtue anyway. Because you are interested in comparing delegate counts, you want the estimate that best captures the delegates effectively determined to date. That includes the second-stage delegates in caucus states. The bottom line is that the delegate count is very close, and there's no "secret Hillary win" that the press is suppressing.

Raised By Republicans said...

Dr. S. you are using the language of sampling and statistical estimation I think. That is, you seem to be arguing that a more conservative operationalization should yield an estimator that is more likely to disconfirm your hypothesis so when it in fact confirms your hypothesis you have a stronger case.

But the problem here is that LTG is right. We will shortly have an actual count, not an estimation but the actual final count of all the delegates. In that event we won't need to estimate anything. Since the various preliminary counts are all over the map, why not just wait for the official numbers to come out?

Or is the goal, rather that you want to give the impression that Hillary is winning? And doing so when the final numbers come out won't be as effective as doing so now.

Finally, on a side note: Given that the New York Times is both Hillary's most recent "home town newspaper" and they've endorsed her candidacy, don't you think you should be a little suspect of their count? Especially if it disagrees with so many other sources?

Dr. Strangelove said...

Basically what I saw yesterday was that NY Times has over 900 pledged delegates for Clinton, something like 70 more than other major sources. Yet they claim to have a very conservative way of counting. So ordinarily one would assume everyone else would have at least what they do, and more liberal estimators would count more. But for Clinton, everyone else had less.

On the other hand, everyone else managed to find all the delegates for Obama that were missing from his total at NY Times. I was aware of the undercount and did my best to compensate in my figures, using the estimates from CNN that considered all pledged delegates tied to the race, no matter the stage.

So I was starting to wonder why the NY Times could find considerably more delegates for Clinton considerably faster with a more conservative count than anyone else, but everyone else could find Obama's extra delegates quickly. I started to wonder if something more than slow accounting might be afoot. Trust me, if you guys noticed the same for Obama, you would have written a post or two!

It is possible, as RbR suggests, that the NY Times has made a mathematical error or has deliberately miscounted. As he and LTG indicated, we will wait for the final delegate count. But if it proves that the NY Times was right that Hillary came out of Super Tuesday with a distinct advantage, I think it is worth considering what happened here in terms of the news narrative and why it occurred like that.

The Law Talking Guy said...

Dr.S., it turns out to be simple. The NYTimes count is not pledged delegates only. It includes superdelegates!!1. I actually added up the NYTimes columns and came to this result. The explainer doesn't mention this clearly. That's the issue, Dr.S. So, the NYTimes is actually showing pledged delegates of Clinton 708, Obama 642. There's no story there. Period.