The only thing I would add (4.00 / 1)

is that the schedule may matter a great deal.

I was in NH this weekend, and someone who knows politics in NH well predicted Iowa would go on Jan 5th, and NH on the 8th.  He was familiar with the work I had done on Iowa's impact of NH, and wondered if the compression in time would increase or reduce the bounce out of Iowa in NH. 

My guess is it would reduce it - which was his feeling as well.

Bottom line - and the Edwards people know this - Edwards needs to get to the high teens or low 20's in NH before the caucuses in Iowa. 

---

by: fladem @ Thu Aug 30, 2007 at 11:56:05 AM CDT

Actually (4.00 / 2)

I'm pretty sure a three day window would increase the impact of the bounce. would increase with only a three day window. Looking back at 2004 New Hampshire polls, Kerry's bounce reached its peak in the polls conducted 3-4 days after Iowa. Since the bounce started to recede in the 4-5 day post Iowa polls, that means the peak of the bounce was 3 days after Iowa. The same pattern can be found in 2000, when both Al gore and George Bush's bounces peaked on January 27th, three days after Iowa. Structurally, it seems like three days would keep the Iowa bounce intact.

The question is whether the news coverage would change in such a window. Post-Iowa coverage would basically only be on Sunday, which would be huge of course, but which would also be short. The shorter post-Iowa news cycle might decrease the size of the bounce. Also, New Hampshire might not receive much coverage at all.

Hard to say, of course. We need to see the calendar before making any guesses.

---

by: Chris Bowers @ Thu Aug 30, 2007 at 12:23:18 PM CDT

[ Parent ]

I wonder ... (0.00 / 0)

... does that timing reflect that it took until three days after Iowa before Gov. Dean started doing serious national tv to rebut the "scream" narrative, rather than anything inherent in the notion of three days post-caucus?

---

by: Adam B. @ Thu Aug 30, 2007 at 17:46:42 PM CDT

[ Parent ]

Fourteen and a half (4.00 / 1)

I love the work you did, quantifying the bounces, but I think we'd be wrong to consider those figures anything close to gospel. They're based on 1988 and 2004 only, yes? Small sample, and I think that to extent that this year's campaign is unique (do you think it is, at all?), those numbers could fall anywhere.

V. interesting, for the poll-obsessed, but with the dates in flux, the campaign so long, the current narratives, and the remarkable weakness of the R field, I'm not sure how much predictive value we're talking about. Oh, and of course Gore's gonna enter.

---

by: Joel @ Thu Aug 30, 2007 at 13:17:12 PM CDT

[ Parent ]

I may not agree with you about Gore but... (4.00 / 1)

I do agree absolutely that we can't put too much trust in the averages. They're ballpark figures, liable to change depending on who wins and how much of a surprise that is.

For example, I'd imagine a Clinton victory in Iowa would provide a bounce in NH, but not a great one, as it wouldn't be perceived as any great surprise. If, on the other hand, she were to finish third (even if it were only a very small way away from being second) the ramifications would probably be greater as it would damage her appearance of being dominant.

Right now Obama probably has the greatest potential bounce, but if polls continue to show Edwards falling back and the national media begins to write him off in Iowa then a late surge and good regional numbers help him to victory, then his bounce could be immense.

Whilst the polling may not be moving especially quickly, it's still close enough and the caucus procedure is abtruse enough for anybody to obtain a potentially nomination-winning or candidacy-ending result from Iowa. And that's without Richardson or one of the other candidates making up extra ground.

Forgotten Countries - a foreign policy-focused blog

---

by: Englishlefty @ Thu Aug 30, 2007 at 18:12:32 PM CDT

[ Parent ]

Wrong (0.00 / 0)

Fladem, you and whoever you spoke to are wrong.  Look at the tracking polls on the unh survey center site for the 2004 primary.  Howard Dean took the first three or four days after Iowa to start turning around after the post Iowa plunge. he ended up a respectable second. Edwards never had the time to build on his second place finish in Iowa.
If Iowa goes three days or less before NH, however, it won't matter too much, because no one will pay attention to NH. I've talked to enough people in campaigns and the national media to know that they are all saying that if NH is three days after Iowa, NH will be pretty much ignored.  Unfortunately, for those of us who love the NH primary, I am afraid that this is something that may happen. 

---

by: Kathy Sullivan @ Thu Aug 30, 2007 at 16:58:13 PM CDT

[ Parent ]

For the record (0.00 / 0)

the person was a county chair in NH.

Here is the way reporting works after Iowa:
Monday - Iowa Caucus
Tuesday - Iowa postmortem.  NH coverage driven off of Iowa. Example, the Concord Monitor's lead story on the FP in 2004 began by describing a race remade.
Weds - Thursday - poll data starts to reflect Iowa impact, which in turn drives coverage.  Candidates not in top 3 are starved for attention.
Friday - Polling finally reflects all of Iowa' impact/  Horse race reporting at its peak.
Sat - Mon - reporting begins to re-focus on the actual race, not the fallout from Iowa.

3 days may not be enough to get the full bounce.  Had the NH primary been held 3 days after Iowa in 1984, Hart would not have won (I know about '84 personally).  I cannot speak about 2004 directly, though I note that Kerry began to moce in NH before Iowa - a reflection of the fact that the Iowa effect on NH begins before the caucuses themselves.

5 days would maximize the bounce out of Iowa, but not give time for the Iowa loser to right ship.  This would be the best case for Edwards.

7 days gives someone enough time to come back from a bad Iowa showing.  They are better examples of this on the GOP side (Reagan's I am paying for this microphone eg), but Dean began to fight back on the Sunday talk shows in 2004.

I don't agree about the bounce out of NH being minimized.  In fact, if you look at the national data, Kerry got about a 15 point bounce in the national polls BEFORE NH but after Iowa.

---

by: fladem @ Thu Aug 30, 2007 at 23:03:38 PM CDT

[ Parent ]

I agree, Obama needs to beat Clinton in IA (0.00 / 0)

to stay in this thing. I wouldn't put it past him--they are building a strong ground game, and we know he will have plenty of volunteers too. The demographics work against him in Iowa, with a lot of precincts containing almost no young people--that's Obama's main problem here.

Edwards needs to improve his standing in NH before Iowa, and then win Iowa, preferably with Clinton in third.

If Iowa caucuses during the first week in January, it is going to be very hard to do GOTV. Only the diehards will show up, in my opinion.

Join the Iowa progressive community at Bleeding Heartland.

---

by: desmoinesdem @ Thu Aug 30, 2007 at 13:04:30 PM CDT

desmoines dem (0.00 / 0)

  I'll put it this way I think Edwards campaign is in serious trouble despite Iowa, unions are not supporting him enough and without them how is he going to compete on FEB 5, no matter who his main "challenger is" even assuming he's seen as in the top 2 on that date. I understand what Elizabeth is doing bringing up "Clinton hatred" and I agree with it, but I thought that would be a more subtle theme,to go that hard on it with dem primary voters is probably suicide. Frankly Obama winning Iowa or significantly outpacing Hillary in the state regardless of what Edwards does is the only way to stop her. New hampshire is not a normal state, independants who have no presence in polling now will determine alot and are very likely to give Obama a real shot especially following a strong Iowa showing. If stopping Hillary is that important to an Iowa caucus goer voting for Obama is by far the best way.

---

by: nevadadem @ Thu Aug 30, 2007 at 13:28:34 PM CDT

[ Parent ]

Two comments... (4.00 / 1)

1.  I can see why Edwards hasn't been tearing into Obama at all.  It's really in the interest of Edwards that Obama stays a credible candidate who can take second place.  It's also in Obama's best interest Edwards stay in the running, though he isn't as dependent.  This runs counter to the expectations of partisans for both candidates that one needs to step aside so the other can be the "anti-Hillary."  Their current tag-team on Hillary really is the best bet. 

2.  Being reminded of the Iowa Caucus structure, I can't help but wonder if Obama's large amount of field offices in Iowa are related.  EG, is part of his strategy to make sure he cleans up big in rural areas, sweeping up those votes who would have showed up for Hillary (or whoever), but doesn't break the needed threshold.

---

by: telephasic @ Thu Aug 30, 2007 at 13:28:50 PM CDT

his strategy is to address his weakness (4.00 / 1)

around older and rural voters so far. The Obama campaign knows that he is strongest among young people and college kids, but they are only found in large numbers in a few counties.

He needs to have extra field offices to address the natural advantages Iowa's demographics give to Edwards and Clinton.

I don't see Obama "cleaning up big" in rural areas. Those were Edwards' strongest places last time.

Plus, Edwards has continued to campaign in smaller towns and has put forward strong rural recovery policy proposals.

Plus, he has the backing of Denise O'Brien, a highly respected organic farmer who was the Democratic nominee for secretary of agriculture.

Join the Iowa progressive community at Bleeding Heartland.

---

by: desmoinesdem @ Thu Aug 30, 2007 at 13:54:16 PM CDT

[ Parent ]

Except (4.00 / 1)

Fladem's numbers are averages that don't take into account other factors. The bounce from Iowa isn't just based on what place the candidates take- the bounce will be different based on Iowa expectations and pre-caucus placing in NH. There is no way the bounce Clinton would get from winning Iowa would be as big as the bounce Edwards or Obama would get. She's the New Hampshire and national frontrunner with universal name recognition. I'm not saying she's hit a ceiling but its hard to pick up momentum if you're already winning. On the flip side any outcome where Clinton doesn't win undermines the inevitability meme. Her drop would be bigger than Obama or Edward's. Further, if the  Inflated Clinton Polling theory is true then the drop she would take from placing third would be even bigger.

The media's expectations narrative in the run-up to the caucus will also play a huge role. If Edwards were to regain a clear lead now and hold it to the caucus the expectation would be an Edwards victory and his bump would be diminished. But if there is a three-way tie going in the he'd get a larger bounce bounce.

Obama has another dynamic going for him. A black candidate winning in Iowa or New Hampshire demonstrates national viability (which I guess mean's "white people liking you" these days). The media would fawn over him for weeks afterward.

---

by: Jack @ Thu Aug 30, 2007 at 14:02:45 PM CDT

I agree (0.00 / 0)

I've been working on a post for TPM Cafe, and in it, I basically argue the same thing: the Iowa bounce is at least in part fed, if not totally created, by media narrative.  If the narrative out of Iowa is that Edwards won, but it wasn't that impressive a win considering the amount of time he's spent campaigning there (years leading up to this race, basically), then I don't think his bounce is that great.  I also think you accurately get how big a hit the "Clinton isn't inevitable!" narrative would be to her campaign.

What's more, if it's very clustered, i.e., say, Edwards/Clinton/Obama 28-25-23, that might end up benefitting Obama, if the narrative that comes out is that Clinton isn't inevitable, but that, in spite of his Iowa win, Edwards isn't nationally viable given his weak support in NH, SC, country-wide polling, etc.  Basically, if Obama is very close to Clinton, he could emerge as the only viable alternative to a suddenly-beatable Hillary even if Edwards wins Iowa, because Edwards is so weak everywhere else.

Now, this is all speculation.  My overall take, however, is similar to Bowers':  Clinton is in strong but certainly not over-whelming position, Obama is well within striking distance, and right now, Edwards has little chance. 

---

by: mopper8 @ Thu Aug 30, 2007 at 15:27:16 PM CDT

[ Parent ]

goot take on it (0.00 / 0)

  ..also E Edwards direct shot on Hillary does show some desperation. THe point she was making requires alot more subtlty with dem primary voters. I hope it helps Obama who knows better than to have his campaign make the vote for me becuase so many people hate Hillary case so explicit.

---

by: nevadadem @ Thu Aug 30, 2007 at 19:29:38 PM CDT

[ Parent ]

The Iowa bounce is a creation of three things (0.00 / 0)

1.  The media narrative.
2.  The effect it has on voter's understanding of a candidate's electbility.
3.  The fact that the lower tier looks much more irrelevant than they did before Iowa.

People hate talking about it, but electability in Iowa and NH matters a low. One impact of Iowa is that the winner starts to look more electable.  This is not as much a media creation as it is a perception voters arrive at on their own.

---

by: fladem @ Thu Aug 30, 2007 at 23:27:53 PM CDT

[ Parent ]

If you read the diary (0.00 / 0)

You would find I couple of things:

1.  I note that media expectations matter.  This was what gave Hart the huge boost in NH ( I know - I was in NH the week after Iowa in '84 working for Hart). 

The lead story out of Iowa will about Clinton.  If Clinton loses, the candidate who beats her may get a bounce like the one Hart got when he beat Mondale (he went up 28 and Mondale went down 10)

The toughest part of predicting 2008 is figuring out what happens to Hillary id she loses.

Flip side - even the biggest front runner has gotten a bounce in the National Numbers out of NH - even Gore did in 2000.

2.  You suggest Edwards would not get a bounce if he was expected to win.  Here the 1988 example is instructive, since Gephardt was expected to win Iowa and still got a 14 point bounce. 

---

by: fladem @ Thu Aug 30, 2007 at 23:21:03 PM CDT

[ Parent ]

I don't (4.00 / 1)

disagree that the winner of Iowa will get a bump no matter what. I'm not critisizing your analysis of the bump in past primaries. Rather, I'm disputing the validity of applying averages of those past bumps uniformally to each candidate in these scenarios. To begin with we don't really have enough samples to predict future bounces with much confidence. Second, assigning the bounces by position in each scenario while ignoring which candidate is in the position is a planely inaccurate way of predicting the bounce since we all admit that each candidate would get a different bounce from winning. 

---

by: Jack @ Fri Aug 31, 2007 at 00:02:13 AM CDT

[ Parent ]

Small and large bounces (4.00 / 1)

I hate to complicate things, especially since I really like the idea of a momentum indicator. But I notice that that historical data (such as it, with only four multi-candidate data points) is basically bimodal (two points at 15, two points at 36), although it does average out to 25. 

My hunch, just from the history of those two elections, is that come from behind victories in either Iowa or New Hampshire (Kerry in '04, Hart in '84) are worth more than double. Doing as expected solidifies people's opinions that everything is going according to plan, leading to a moderate but not really necessary bump, but having one candidate do much better or worse sort of resets the clock. Hard to say now which victories would count as come-from-behind victories, but I think it has the potential to really change the equation.

---

by: brklyngrl @ Thu Aug 30, 2007 at 14:15:33 PM CDT

The key distinction (0.00 / 0)

which I make in my diary is whether the front runner wins or not.

In two cases clear front runners lost - and what is astonishing is the overall swing from front runner to the candidate who won was almost identical 49 points in 1984 and 48 points in 2004. 

A front runner who wins jump is more modest, but since they start at a high level it usually ends the race.

---

by: fladem @ Thu Aug 30, 2007 at 23:31:56 PM CDT

[ Parent ]

Inflated Clinton poll theory IN IOWA? (0.00 / 0)

Mark Blumenthal talks about the difference between the two Iowa polls released today.  Good stuff, read it all here.

Key graph:

The resolution of that argument is neither simple nor obvious, but seems to have a profound impact on the results. Surveys that appear to include more past caucus goers (Time, Des Moines Register and One Campaign survey -- see our Iowa compilation) tend to favor John Edwards, while Hillary Clinton does better on surveys that define the likely caucus-goer universe more broadly.

---

by: mopper8 @ Thu Aug 30, 2007 at 15:33:10 PM CDT

As soon as the Big Dawg... (4.00 / 1)

...starts campaigning in earnest for his wife, I think you can put this one to bed.

...It is sort of like gravity.  Bill Clinton's aura is a huge enough draw to suck in all the free-range Democrats.  I just hope the hard won progressive direction of the modern Democratic party can be somehow maintained.  How that happens when the DLC is once again in the driver's seat I do not know.

---

by: Deep Blue @ Thu Aug 30, 2007 at 16:32:55 PM CDT

he's campaigned for her in Iowa (0.00 / 0)

and its had very little impact

---

by: bruhrabbit @ Fri Aug 31, 2007 at 09:31:44 AM CDT

[ Parent ]

Asked this before, but no one really answered (0.00 / 0)

Clinton is only ever receiving in the 30s in the polls. Even in NH. She has universal name ID. She has strong likeability. She has a lot of money advantge etc. Why exactly is she only obtaining this number? Is it her ceiling rather than her floor of support? And, what does this mean as other candidates fall out of the race such as Kucinich, etc?

---

by: bruhrabbit @ Fri Aug 31, 2007 at 09:33:53 AM CDT