Update, January 5, 2008: With Obama and Huckabee winning Iowa - both parties' nomination battles are up in the air. Accordingly, I thought it appropriate to re-offer this essay, which was originally published on December 30, 2007. In what follows, I review how momentum works on a statewide basis, and try to piece together what big primary victories mean for candidates later on. In light of this week's Iowa results, I hope you will find it worthwhile.

***

On Friday, I offered the first of a two-part series on momentum. I examined momentum on an individual level, reviewing how voters may be influenced by it. Today, I will discuss whether and how candidates with momentum can win primary and caucus contests.

There are two broad ways to characterize statewide momentum - though, practically speaking, momentum in any given year probably works through a combination of both ways. First, it can build slowly - so that a win in New Hampshire yields a win in South Carolina, which in turn yields a win in Florida, and so on. This would be like a truck rolling downhill, slowly gaining speed. Let's call this piecemeal momentum. The other way it can work is what we'll call all-at-once momentum. A candidate wins Iowa and/or New Hampshire, and everything follows from that dramatic beginning. This would be less like a rolling truck and more like a space shuttle that uses the launch pad for blast off.

At first glance, piecemeal and all-at-once momentum would look the same - but in fact the causal processes would be different. With piecemeal momentum, a win in New Hampshire would influence South Carolina, which would influence Florida. With all-at-once, a win in New Hampshire would influence South Carolina and Florida, which would really not influence one another. Again, these are broad categories that probably cannot capture specific instances of momentum perfectly. They are just meant to help us frame the conversation. In reality, we might expect a candidate's momentum to be a combination of both - perhaps a huge launch in New Hampshire and Iowa to which subsequent victories add a little bit every time.

Let's start our investigation with piecemeal momentum. Researchers have found it does exist - some candidates do build momentum slowly over time, victory-by-victory. However, Arizona's Barbara Norrander has found that this type of momentum more usually characterizes also-rans rather than nominees. Norrander tracked election returns in all fifty states between 1976 and 1988, and found that those who eventually lost nominations - like Reagan in 1976 or Hart in 1984 - often made use of this momentum. The eventual nominees usually did not; rather, their state-by-state performances were more frequently conditioned by how much money they spent. The exception was George H.W. Bush, who enjoyed the "Big Mo" in 1988 (even though he thought he had it in 1980!). While Norrander's study is somewhat dated, my intuition is that the pattern still holds. Bush '92, Clinton '92, Bush '00, and Kerry '04 did not win by slowly building momentum.

This is not a huge surprise. On Friday, we discussed the importance of the time between contests. For most of the modern era, primaries have been scheduled quite close together. This probably has limited the effect of piecemeal momentum. Norrander and others have shown that nominees win by spending money, and that early victories can bring dollars to cash-starved candidates. This provides an opening for piecemeal momentum to work: an insurgent candidate wins a big victory that he can build on because donors fill his coffers with cash. Indeed, this was one way Jimmy Carter won his party's nomination in 1976. The trouble is that - for most of this era - the next round of contests always comes too quickly for sufficient funds to be raised. Sooner rather than later, the insurgent who won an unlikely victory runs into the frontrunner's expensive firewall.

So, we might infer that, because of this year's hyper-compacted schedule, piecemeal momentum will not be decisive. However, I think such a conclusion is hasty. The case of the 2000 GOP nomination gives me pause. During the 2000 Republican nomination battle, Penn's Annenberg Public Policy Center conducted a rolling survey of New Hampshire, South Carolina, Michigan, and the Super Tuesday states. Its results are worth considering.

Bush defeated McCain in the South Carolina primary held on February 19th, 18 days after the New Hampshire primary. Interestingly, when Annenberg began its rolling cross-section of South Carolina, McCain's position in the Palmetto state was about as solid as his position in New Hampshire was on the day of his victory. His problem was that his standing eroded as time wore on. But if South Carolina was scheduled as close to New Hampshire in 2000 as it is in 2008 - 11 days instead of 18 - McCain's standing might have been strong enough to win. A win there, plus his wins in Michigan and Arizona, might have sustained the post-New Hampshire boost he registered in the Super Tuesday states.

In other words, McCain might have been just a few days away from pulling off the rare feat of riding piecemeal momentum to the nomination. And so, we cannot disqualify it from being a factor in 2008. Although, we must admit that it is unlikely. Above all, McCain's campaign in 2000 was based upon a huge, 19-point upset in New Hampshire. It was lightning in a bottle. Even if the schedule is more accommodating this cycle, lightning in a bottle is still hard to capture.

So much for piecemeal momentum. What about all-at-once momentum? This, as I said, is like a space ship launch. The idea is that a candidate wins an early state and is helped for the rest of the way.

For a long while, political scientists have known that you can do a pretty good job predicting the outcome of a nomination battle based on a few basic variables, all of which are known by the end of the pre-election year: national poll position, money raised, television exposure, endorsements, etc. Recently, however, scholars have found that when you add the results of the New Hampshire primary to these models - the predictive power jumps through the roof, from around 60% to upwards of 90% (though the effect is muted when we look at just the Republicans). The implication here is that the second type of momentum - the big win that launches a candidate - is a factor.

But let's not get ahead of ourselves. These models predict final shares of the vote - not whether a candidate wins or not. If you look at matters more closely, you see that what New Hampshire has most frequently done is the following. If it backs the frontrunner, he has an easy path to victory - as with Gore in 2000. If it does not back him, he has a harder path - as with Mondale in 1984. The only exception to this rule is Howard Dean in 2004. His loss in New Hampshire (and Iowa, which we will discuss presently) effectively ended his candidacy.

Of course, Dean was a poor frontrunner by many metrics. He wasn't well known. He wasn't well liked. Voters were concerned he was unelectable. The scream. And so on. Dean was easily one of the weakest frontrunners in the modern era.

The fact that he went down in New Hampshire actually does a great deal to clarify the role that the Granite State can play. Clearly, it is a force that is independent of the factors that make up the pre-primary frontrunner (strong national polling, money, etc). When it works in concert with those factors, supporting the frontrunner, it adds to his power. When it works against those factors, supporting an insurgent, it serves as a countervailing force. Usually, the frontrunner has so much power behind him that he can overcome a loss in New Hampshire. But, when a frontrunner is not very powerful - as Dean was not - New Hampshire can stop him.

What about Iowa? Researchers have found that Iowa does not exert the independent force that New Hampshire does. This is how DePaul's Wayne Steger characterized the Iowa caucuses this year in Political Research Quarterly:

The results of the Iowa Caucuses do not significantly affect the prediction of the aggregate primary vote...That the Iowa Caucuses played a crucial role for Jimmy Carter (1976) and John Kerry (2004) is more than offset by the fact that the eventual nominees of both parties have "lost" the Iowa Caucuses in most other years. The New Hampshire primaries, in contrast, have a more consistent effect across nomination campaigns - inflating the chances of candidates who do well while deflating the remaining primary vote shares of candidates who do not.

So, Iowa does not have a direct effect. But, it could affect New Hampshire, which in turn influences the rest of the nation. To investigate this hypothesis, let's consider the following chart, which reviews Iowa, New Hampshire, and nomination victors in contested battles since 1976:

Two observations are relevant. First, Iowa and New Hampshire frequently disagree. They only have agreed four times, which is a sign that Iowa does not usually influence New Hampshire. Two of those four times, 1980 D and 2000 D, were contests where there was a clear frontrunner, so it probably would not be fair to say that Iowa influenced New Hampshire. The other two times were 1976 D and 2004 D. Here, you could make a case that Iowa influenced New Hampshire. And what did those two elections have in common? They both lacked strong frontrunners.

Thus, the influence of Iowa and New Hampshire seems to depend upon the quality of the frontrunners. Iowa can influence New Hampshire, and New Hampshire can upend the campaign dynamic, when the frontrunners are not strong. When the frontrunners are strong - at best, they only slow the march to victory.

This gives us an angle on understanding the role momentum might play this cycle. Clearly, the Republicans have no pre-election year frontrunner - like the Democrats in 1988. This means that momentum definitely could be a factor. As I said, we probably will not see the kind of successful slow-building momentum akin to what McCain almost had in 2000, though it is still possible. What is more likely is momentum that comes from a win in Iowa and/or New Hampshire - a candidate then uses those victories to launch himself beyond the rest of a lackluster field.

What about the Democrats? Over the summer, the press had built Hillary Clinton up as inevitable. Now, it sees her as particularly vulnerable. The reality has always been in the middle. She is a roughly average frontrunner with real strengths and real weaknesses - it will thus be hard, but not impossible, to beat her. Clinton's lead in the national polls right now is of middling size relative to past frontrunners: larger than what some frontrunners enjoyed, smaller than what others enjoyed. This is good for her. Furthermore, her supporters tend to be a good sample of the traditional bases of Democratic support. This is also good. But she has weaknesses. She has not dominated fundraising. If you factor out her Senate campaign transfers - you'd have to say that Obama out-raised her (without using federal PAC money). Further, Clinton's negative ratings are higher than previous frontrunners'. As frontrunners go, she is certainly stronger than Dean. But she is weaker than several past frontrunners. She is far from immune to being seriously affected by momentum. Obama could be a particular threat because he - unlike other insurgent candidates - can fight her dollar-for-dollar.

All in all, it is impossible to assign odds on whether momentum will be a factor this cycle. But there are strong indications that it could be. The bottom line is that neither party has a decisive frontrunner - which opens an opportunity for momentum to play a determinative role in who wins these battles.

Posted by Jay Cost at 08:00 AM | Permalink

With the New Year upon us, the pre-election campaign is ending, and a new phase of the cycle is beginning. Candidates will battle one another from Iowa, to New Hampshire and South Carolina, and eventually to Super Tuesday and beyond. Momentum could be the critical factor - a great victory could propel one of the hopefuls beyond the rest of the pack.

For as important as it might be, people seem to know very little about momentum. The word itself is just a metaphor, and the concept it references is often understood in vague and even mystical ways - a candidate wins a primary and momentum "magically" sends him or her to the nomination.

Over the course of the next two days, I hope to offer some clarity and precision on the subject. Today, I will look at what momentum is when it occurs. Monday, I will look at the conditions in which it may occur.

Let's start with a simple definition that we can build on. A candidate has "momentum" when his or her previous electoral results influence subsequent ones. The reference to Newtonian physics actually fits pretty well. Imagine a truck rolling down hill - and you can get a pretty good sense of the "big mo." Of course, momentum can work in the opposite direction. A win can yield more wins, but a loss can yield more losses.

We usually think of momentum from a "macro" perspective. For instance, we imagine the New Hampshire returns influencing the South Carolina returns. But momentum actually has its origins in the minds of individuals. So, our investigation of momentum should begin with them. In particular, we should first ask about the context of their decision-making processes. What are the factors that might make momentum so influential in their thinking?

As we have discussed all year, this context is dominated by voter inattention and ignorance. The average voter pays little attention to politics, and so has little knowledge of it. This is how momentum can have such an effect. An electoral victory is big news to a voter who knows relatively little about the race. He undoubtedly hears about it, and so hears lots of positive information about the winner. As he did not know much to begin with, this information can be critical to his decision-making. Unsurprisingly, researchers have found that momentum can have its greatest effect on those who do not pay as much attention to the campaign.

Another important contextual factor is the media. After all, it is not just the results, but also the way they are interpreted, that generates momentum. It is one thing to read, "McCain, 115,490; Bush, 72,262." It is quite another to see, "MCCAIN UPSETS IN GRANITE STATE / Trounces Bush By 19 Points." Perhaps the biggest beneficiaries of the press are the surprise victors. The media loves electoral surprises - and they cover successful insurgents as though they walk on water.

Another factor is the schedule. If all states voted at once, there could be no momentum as no voter would know about any prior results. On the other hand, if there were one primary every six months - there would be no momentum, either. The influence of the previous contest would dissipate before the next context. Because momentum is a temporary effect, the timing of the contests matters.

With this context in mind, how can a voter be influenced by prior election returns?

Political scientists have identified several ways that momentum can work. The operative word here is "can." Momentum is a difficult concept to study. Good data is hard to find. For starters, elections in which momentum is a major factor are rare. Another problem is that you need a special type of data to study momentum thoroughly. Simple horse race polls will not do. You either need "panel data" or a "rolling cross-section" that investigates voter psychology as the campaign progresses. These are hard to come by. So, we cannot confirm how momentum really works. At best, we can see whether the data is consistent with various types of momentum.

Five such types have some evidence to support them. The first is "strategic voting." Here, a voter switches his vote from his first preference to a latter one based on who appears to have the best chance of winning. Take a simple example. Suppose a Republican prefers candidates in this order: Forbes, McCain, Bush, Dole, Bauer. He sees that Forbes placed poorly in Iowa and New Hampshire - and that McCain won the Granite State. He decides that because Forbes has no chance, he should support his second choice. Accordingly, McCain's victory in New Hampshire has helped him win a voter he would not otherwise have won.

Another type of momentum is what Princeton's Larry Bartels calls "cue-taking." Voters, in a cue-taking scenario, lack sufficient information to know which candidate is optimal. Previous electoral results serve as signals from like-minded partisans about the candidates. Cue-taking is like heeding advice from friends, neighbors or co-workers when you can't decide for yourself. It is a way to make a choice amidst uncertainty.

The concept of "contagion" (again borrowing Bartels' language) is related to cue-taking, but it is less conscious. Instead of purposefully following the signals of previous electorates, a voter switches his preference just because the victorious candidate is mentioned more often and more positively. This type of momentum is less thoughtful than cue-taking. Bartels compares it to the spreading of a communicable disease. Voters acquire an inclination to support the candidate because those around them have it.

A fourth type of momentum is the pleasure one derives from supporting a winner. We all like to know that our favorite candidate turned out to be the winner - some voters may even switch their votes to get behind the expected winner.

These four forms of momentum were covered in detail by Larry Bartels, whose Presidential Primaries and the Dynamics of Public Choice remains a core text in the study of the modern presidential campaign. Arizona State's Patrick Kenney and Iowa's Tom W. Rice suggested a fifth type - which they call "inevitability." Here, voters support a candidate because he or she is bound to win. The difference between this and the previous form of momentum is that this does not carry the same psychological satisfaction. You back the inevitable candidate not to get the pleasure of supporting the winner, but because there is no other viable choice.

Again, it is difficult to confirm whether and how any of these forms of momentum operate on the individual voter. The data is just too limited. Nevertheless, Bartels found that momentum does not need to have a direct effect. A voter need not have his vote choice directly influenced by the latest electoral returns. Instead, the returns can influence his evaluations of each candidate, which then influence his vote choice. This would be a process like projection - in which a voter comes to believe that the advantaged candidate possesses all of the qualities he most wants in a leader. This positive reevaluation thus makes the candidate the "obvious" choice for the voter's support. Meanwhile, Penn's Diana Mutz found that momentum can work similar to the way that one person influences another: when a voter learns that an electorate voted differently than he would have, he is induced to reconsider his views and, "think of arguments that might explain those others' positions. By rehearsing these arguments, people engage in a process of self-persuasion whereby their own attitudes move in the direction that have been primed by others' views."

These considerations yield several relevant points for 2008.

First, the media matters. Over the course of this cycle, I have distinguished between the perpetual campaign and the real campaign. For most of the year, the media has been captivated by a drama that average voters barely noticed. This is the perpetual campaign - the endless struggle of the candidates to win the news cycle. It has little effect on vote choice - though pundits often think it does. The real campaign starts when candidates communicate directly with voters, who begin to make up their minds. During this campaign, the media can have a real influence because it interprets electoral returns for voters. The tone of the interpretation can determine whether a candidate will acquire momentum. Suppose, for instance, that Romney finishes second in Iowa. Will it damage his prospects in subsequent contests? It might. One of the factors that will matter is how the media covers it. Was Romney "stunned and decimated, spending millions only to be defeated by the insurgent Huckabee"? Or did Romney, "rally and win a good chunk of evangelicals, despite the impediment that his Mormonism presented"? Both storylines could be consistent with the Iowa results. Which one the press chooses could influence what happens next.

Second, media expectations matter - and candidates will try to manage them. The Huckabee campaign, for instance, does not want the media believing he is assured an Iowa victory. If he wins "as expected," the press might not cover it as glowingly - thus diminishing his bounce. So, Huckabee is out there saying his goal is to finish in the top three. This is why Bill Clinton said he would be amazed if Hillary Clinton wins Iowa. He knows as well as anybody that expectations matter. His second place finish in New Hampshire in 1992 effectively became a victory because of the press coverage.

Third, voters' second choices matter. They definitely matter to Iowa Democrats, who are sometimes forced to compromise at the caucus meeting. But they also matter to voters all across the country - as first choices seem like they can't win, strategic voting might induce voters to fall back on their second choices. The only poll I am aware of that asks about second choices is Cook/RT Strategies, which is currently out-of-date. But you can get some purchase on second choices in other ways. Examine candidates' favorable/unfavorable ratings - those with higher favorables are more likely to be second choices. Also, look at which candidates win on salient personal qualities (trustworthiness, for instance). This can give you a sense of who might be the second choice of a primary electorate.

Fourth, momentum can help some candidates but not others. As mentioned above, the level of information that voters possess matters when it comes to momentum. So, it stands to reason that candidates about whom a lot is known will not benefit from momentum as much. Walter Mondale received no bounce after he won the Iowa caucus in 1984. As a former vice-president, he was well known by the voters. On the other hand, Gary Hart enjoyed a huge bump after he won the New Hampshire primary because voters did not know a lot about him. So - the high profile candidates this season stand to benefit less from big wins. The low profile candidates might get a real boost. Accordingly, you should note the electorate's inability to rate various candidates in the pre-election polls. This is a sign of how much they know about them - and whether they are susceptible to momentum.

Finally, remember that our data limitations make momentum an under-studied concept. It is thus hard to predict when and where it might make an appearance. Those who talk confidently about momentum are actually just speculating.

Posted by Jay Cost at 12:00 AM | Permalink

George Mason University's Center for Media and Public Affairs published the major findings of a study of media coverage of the presidential candidates. It examines the nightly news on CBS, ABC, NBC, and Fox (the first half hour of Special Report) from October 1 to December 15.

The results are interesting and consistent with what we generally know about media coverage.

First, the press gave both parties' insurgent candidates - Obama, Edwards, and Huckabee - more positive coverage than the frontrunners. Below is the percentage of positive coverage for all of the major candidates:

McCain: 33%
Giuliani: 39%
Romney: 40%
Clinton: 42%
Thompson: 44%
Huckabee: 50%
Obama: 61%
Edwards: 67%

Most of the insurgent candidates are all grouped at the end of the list. The only exception is McCain - whose negative coverage might be explained by the fact that his campaign belly-flopped in the summer. Generally, this ordering is consistent with what we know about media coverage: it tends to be more generous to the candidates who are behind, and harsher to the candidates who are ahead.

The report also finds that the emphasis of the news coverage is on the horse race and not substance. It finds that:

122 covered candidates' personal backgrounds
188 stories covered policy issues
191 covered campaign tactics
162 covered candidates' standings in the race

Only the first two types of stories have a real effect on vote choices - and they are outnumbered by the latter two. However, I am surprised by the roughly even split between the two groups (310 on substance and personality vs. 353 on the horse race ). While the campaign coverage has not been nearly as substantive as it could be, these numbers are more balanced than I would have guessed.

Finally, the report finds that the Republicans were covered more negatively than the Democrats. Specifically, "On the three broadcast networks, opinion on Democratic candidates split 47% positive vs. 53% negative, while evaluations of Republicans were more negative - 40% positive vs. 60% negative" Conservatives might explain this via recourse to anti-Republican media bias - but there might be other explanations as well. For instance, everybody seems to agree that the Democrats are in the better position heading into 2008. Perhaps the bearish orientation to the GOP has induced more negativity in press coverage.

The exception was Special Report with Brit Hume - which came out to be well balanced. The report finds, "[E]valuations of all Democratic candidates combined were split almost evenly - 51% positive vs. 49% negative, as were all evaluations of GOP candidates - 49% positive vs. 51% negative, producing a perfectly balanced 50-50 split for all candidates of both parties. " Interestingly, this is not the first time Hume's show has been found to be even-handed. A study in the December 2005 issue of the Quarterly Journal of Economics by UCLA's Tim Groseclose and Missouri's Jeffrey Milyo found that, by their metric, Special Report was one of the fairest shows on television.

Posted by Jay Cost at 12:48 PM | Permalink

Anybody who checked Drudge today will have seen that there is a "shock poll" that puts Hillary Clinton 15 points in front of Barack Obama in Iowa. The polling company that produced the poll is ARG, and this is what it had to say about its results:

Hillary Clinton leads Barack Obama among women 38% to 21%, which is unchanged from a week ago (Clinton 36%, Obama 23% among women). Obama has lost ground among men to John Edwards and Clinton. Among men, Clinton is at 28%, Edwards is at 27%, Obama is at 16%, and Joe Biden is at 11%. A week ago, Obama was at 27% among men, followed by 21% for Clinton and 19% for Edwards.

This poll might indeed be a trend - the first sign of a swing back to Clinton among Iowa Democrats. Unfortunately, we will not be able to know for a few days - as polling companies presumably suspended operations over Christmas. I have a few caveats that I would put in place on this poll - just a few basic warnings about why we should not over-interpret these results.

***

1. ARG polled the weekend before Christmas, from 12/20 to 12/23. This might not be the best time to construct a sample of likely Iowa voters. No other poll I know of has come out with a sample taken from those days. This is a sign that most other pollsters were wary about Christmas weekend.

2. The ARG poll has Clinton up and Obama down by statistically significant amounts relatively to its last poll (12/16 to 12/20). On the Republican side, it has Mike Huckabee down and Ron Paul up by statistically significant amounts. This is a lot of movement - four candidates made statistically significant moves in the course of three days. Recall the last point, and note that these are three days when respondents probably were not thinking much about politics. December 20th to the 23rd are days usually filled-to-the-gills with last-minute holiday preparations. They are not great days for reflecting on the state of the presidential campaign. Thus, this movement might be due to the sampling effects mentioned in Point 1.

3. There are other elements of the poll that just don't scan with me. For instance, it shows Fred Thompson at 3% and Alan Keyes and Duncan Hunter both at 2%. ARG has shown Thompson low over the last few weeks - so this would not be a consequence of ARG's internal sampling method thrown off by the holiday weekend. But its last two samples estimated Thompson's support well below the rest of the Iowa polls. And 3% just does not pass the "smell test."

4. Mark Blumenthal has noted several interesting facts about ARG. First, they sample more heavily than any poll from first time Iowa caucus goers (on the Democratic side). This is probably why they usually have Edwards below where he is in the RCP Iowa Democratic average. Edwards is doing relatively well among previous caucus goers, but ARG is "diluting" their influence with the first-timers. Now, ARG's intuitions about first time caucus goers may be correct, but they are on the margins on this issue. Second, they did an extremely poor job of reporting their sampling methodology when Blumenthal requested it. They would not provide any information about respondent demographics, and they would not provide information about the number of long-time caucus goers in their Republican samples.

5. Just because differences between polls are statistically significant does not mean that they are necessarily caused by changes in the population. Clinton, Obama, Huckabee, and Paul have all made statistically significant moves - but some of these could still be statistical blips induced by the sample. This is as good a time as any to review exactly what statistical significance is.

The technical language that describes the margin of error usually reads something like this: "We are 95% confident that the true values are +/- 3%."

This is referring to Type I error, or the error of the false positive. It means that 95% of the time, when you take a poll and get 17%, the real world value will be between 14% and 20%. This also means that 5% of the time (or one time out of 20), it will be outside this range. This is the poll's tolerance of Type I error. The chances are 5% that you will have a false positive - you will believe that the real value is between 14 and 20 when in fact it is not.

But suppose you have 20 different statistics you are looking at. What are the chances that the real world value of at least one of them will be outside the margin of error simply due to sample effects? It is 64%!

This is something that is rarely noted when looking at poll trends - it is called the experiment-wise Type I error rate. When you look at the polls to divine trends, you are implicitly doing some form of statistical hypothesis testing. You are trying to determine whether changes are due to sampling error, or whether they are due to shifts in the population. To do this, you have to assume that sampling error will only explain so much variation in the polls. Usually (95% of the time), this assumption holds up. Occasionally (5% of the time), it does not. When it does not hold, you have committed Type I error. And the more polls you look at, the more likely it is that you have committed it.

***

Casual readers, please take note: I am not predicting that this is a blip. Contrary to what some have assumed, I do not make predictions about the ways the polls will move. That is a fool's errand. My point here is simply that it is possible that this movement is induced by sampling effects - and we should be careful not to over-interpret these results.

Posted by Jay Cost at 10:27 AM | Permalink

Most neutral observers would agree that Hillary Clinton's response to Barack Obama's rise has been bungled. Over the past few weeks, we have seen her campaign attempt again and again to attack him, only to make itself look foolish. I think the worst moment came last weekend when President Clinton was dispatched to the Charlie Rose Show to trash the junior senator from Illinois. That task was simply beneath a former president. And who did not notice the irony of Clinton arguing for experience over freshness? If any Democrat has parroted Republican talking points this cycle - it was Bill Clinton mimicking Bush-Quayle '92.

This plan was clearly put together on a spit and a prayer. It seems to me that if the Clinton campaign had anticipated that Obama would pose this kind of threat - it would have developed a better strategy for dealing with him. Its ineptitude over the last few weeks belies its lack of preparedness. I am sure that Team Clinton has a number of contingency plans in its filing drawer, but the rise of Obama is clearly not one of them.

Why was the Clinton campaign unprepared for this?

Unfortunately, we cannot answer this question directly. The only people who know are the higher-ups of the Clinton organization - and they are not going to admit that they were unprepared, let alone explain why. But I have a plausible theory worth sharing.

The way to approach the question is first to ask why we should have expected an Obama surge. It stands to reason that the Clinton campaign failed to account for at least one of the factors that make up our answer. These are the three reasons that I argued for over the summer and fall:

(1) Obama raised $70 million in nine months from half a million people. This demonstrates two points:

(a) He caters to a real demand in the Democratic electorate - intense enough to open wallets.

(b) His money can facilitate a more sophisticated campaign strategy. Obama can do more than win Iowa and hope that he magically catches fire. Instead, he can win Iowa and fight Clinton dollar-for-dollar, state-for-state.

(2) Obama is the most authentic change candidate among the top three Democrats. Hillary Clinton is not this candidate. Her principal qualification for the job is her role as her husband's advisor - so she was always going to run on the record of the 1990s. John Edwards has positioned himself as a change candidate, but he does not convey the authenticity that Obama does.

(3) Obama is organized in Iowa. He recognized that organization was critically important for an Iowa victory - and that an Iowa victory was necessary for his broader strategy. And so, he is organized and ready for the January 3 caucus.

Through the summer and the fall, journalists underestimated the importance of these because they used the opinion polls to create a horse race out of whole cloth. In reality - the opinions expressed to pollsters were not stable enough to support the idea that there was an actual race going on. Voter opinions were based on little information and even less interest in the campaign. Obama's activities were never going to register with these uninformed and uninterested voters in the summer; they were always meant to yield dividends in the winter. So, Obama was seen to be a weaker candidate than he really was. Accordingly, Clinton was seen to be stronger than she really was. She was always the frontrunner (she still is), but the overuse of opinion polls made her appear "unstoppable" and "inevitable" to the press.

Like the press, the Clinton campaign clearly underestimated Obama - it over-looked the money, the message, or the organizing. Perhaps the Clinton campaign did this for the same reason as the press. Perhaps it relied so heavily on the opinion polls that it could not see that Obama was preparing to launch a viable campaign later on.

I think this explanation has some credibility to it. I'm thinking in particular of Mark Penn, Clinton's chief strategist and pollster. His comments over the course of the campaign have struck me as utilizing the same erroneous assumptions that informed the press' summer horse race narrative. Consider this snippet from the Ben Smith's blog. The date of this entry is October 18. 2007:

"Republicans are not prepared for the loss of a substantial group of Republican women voters ... even in the South," he said. "I think you're going to see as much as 24% of Republican women defect and make a major difference nationwide in terms of, I think, the emotional element of potentially having the first woman nominee. And that that actually will be a major unexpected factor here that will throw the Republicans for a loop."

This is a ridiculously overconfident assertion.

First, research has shown that partisanship is a stable and powerful feature of a person's psychology. It has also shown that voters who are conflicted between their partisanship and their evaluations of the candidates often resolve the conflict by simply abstaining, rather than voting for the other party. The idea that one in four Republican women will defy these regularities is possible, but far from likely.

Second, I just cannot see how this figure can be quantified some twelve-and-a-half months prior to the election. That's just insane to me. Everything we know about the levels of voter information, voter attention, the effect of the media dialogue on transitory political opinions, the influence of question wording and ordering - leads me to suspect that Penn is committing some serious inferential fallacy. It is not that he is necessarily wrong. It is that his assertion is dramatically underdetermined. The numbers are not giving him the hard answers he thinks they are.

This is the kind of comment I expect from non-expert journalists who look at the polls and read the numbers in a naïve way. "Republican women say that they would consider voting for Clinton; ergo, they would consider voting for Clinton." Not quite. The fact of the matter is that you can't come to this conclusion so easily. Underlying all of those seemingly straightforward numbers is a complex, intricate aggregation of individual voter psychologies. This makes inferential analysis extremely difficult.

I would contrast Penn's silly assertion with the considered work of political scientists who specialize in political psychology. The best work in this subfield is the most difficult stuff I have ever read. The theories are complicated, the methods are complex, and the conclusions are always narrow and tentative because voter psychology is incredibly difficult to delineate. It is not made any easier by the fact that our best point of contact with voters is the opinion survey - which, when you think about it, is quite distant from their interior mental states.

You can find the same flippancy in Penn's strategy memos - which have come out periodically over the course of the campaign. All of them follow the same basic script as this one from July: Clinton's poll position is insurmountable; there is no need to have an election because a sample of the voting population has reported statistically significant results that Clinton will win.

In January he argued, "If Hillary leads in Ohio at this point in the race -- the key state that gave the last election to the Republicans -- then this confirms that Hillary can win and is today winning. She is the strongest Democrat in what was the most difficult state." Look at his words carefully: polling results some twenty-two months before the election "confirm" that Hillary can win and "is" winning. In February he said basically the same thing in response to the latest polling data, "As other candidates are getting more and more attention, Hillary is getting more and more support...This poll confirms that Hillary not only can win but actually is today winning."

In August, he wrote that voters had "come to see the race differently," and accordingly "concluded" that Clinton "has what it takes to be President and what it takes to take on the Republicans." It is untenable to argue that voter perceptions were changing in August, or that voters had concluded anything before Labor Day.

In October, he said that Clinton's support among women is "deep" - and that "94% of young women" will be more likely to turn out to vote for the first female nominee. Much like his comment about Republican women - there is no way to quantify how "young women" will respond to Clinton some thirteen months before Election Day. Nor, for that matter, can depth of support be easily interpreted from simple yes-or-no questions asked months before the first votes are cast.

In November, he wrote that the "leadership card" is the reason "why people are voting for Hillary Clinton." His consistent use of the present tense to describe the act of voting represents the same fallacy that polling respondents are no different than voters.

Now, there is a lot of spin going on with these memos. In particular, Penn has an incentive to play to the prior beliefs of the audience - i.e. journalists who cannot distinguish between a February poll and a December poll. However, spin alone cannot explain Penn's voluminous study of the polls. I counted upwards of 6,000 words offered in analysis of polling data since the start of the campaign - no campaign's senior adviser spends so much effort dissecting the polls simply for the purpose of spinning the press. That would be a misallocation of resources.

Spin alone also does not explain why Penn would consistently take the tone he has adopted. If he agreed that the polls could turn on a dime, he would never argue that they would not. Anybody with a lick of sense knows not to tell a falsifiable lie if it can be avoided. Generally, the fact that the Clinton campaign facilitated the idea of inevitability is an indication that it believed that it was inevitable. You don't go pushing that storyline if you believe it is not necessarily true. Otherwise, you set expectations far too high - and you run the risk of getting burned when things do not go your way.

This is actually Clinton's biggest problem right now. When the voting starts, expectations matter. Most of the electorate's information on the campaign will come from the media - whose emphasis is on the horse race. Who's up and who's down. If Clinton loses several of the first contests, she will receive more negative coverage than, say, Giuliani because her campaign once convinced the press that she couldn't lose.

So, I think that the Mark Penn and the Clinton campaign might have made the same basic mistake the press made. They over-interpreted the polls. They wrongly assumed that the opinions expressed in them were more informed, sophisticated, and stable than they actually were. I think that this, in turn, caused them to overlook Obama's real strengths.

The big question: what will this cost them? Possibly nothing. I think this race remains fundamentally unchanged. Winning Iowa is a necessary condition for Obama. It is a sufficient condition for Clinton. Clinton could win Iowa - and her numbers that have slipped in New Hampshire, South Carolina, and nationwide will rebound. The race will be over. In that case, her campaign will probably be in better shape for having learned a good lesson at no cost. But it might cost them something. There is such a small margin that separates Clinton, Obama, and Edwards in Iowa that her campaign's maladroit response to Obama could cost her victory. In that case - the race continues to New Hampshire, South Carolina, and then Super Tuesday. And she could lose the nomination.

Posted by Jay Cost at 12:00 AM | Permalink

On Friday, I wrote a column entitled "On the Republican Itinerary." Using data from the Washington Post, I tracked where Republican candidates have been spending their time since November 1 (outside debates and fundraisers). Following the Post, I noted that Fred Thompson had 14 public events, half of which had been in South Carolina.

A few hours after the article had gone up, I received an email from Paul in New Jersey - pointing me to an article written by Bill Theobald of the Tennessean, which asserted:

A review of the two-week period from Nov. 4-17 by Gannett News Service shows Thompson held 28 campaign events, making him No. 2 among the four major GOP candidates. Sen. John McCain of Arizona topped the list at 35. Former Massachusetts Gov. Mitt Romney held 22 and former New York Mayor Rudy Giuliani, 20.

Giuliani, however, had the broadest reach, visiting 10 states and Washington, D.C., during the period. He was followed by McCain with 10 states, Thompson with seven states and D.C., and Romney with six states.

How to explain the differences? Remember first of all that I excluded debates and fundraisers. But also WaPo says about its count, "This database does not include every event; fundraisers often are not announced, and some events will be updated as more details become available."

When we put the fundraisers and debates back into the WaPo count, we can make the following comparisons:

Giuliani: Gannett says 20 events in 10 states plus DC; WaPo says 18 events in 8 states plus DC.

McCain: Gannett says 35 events in 10 states. WaPo says 39 events in 8 states plus DC.

Romney: Gannett says 22 events in 6 states. WaPo says 36 events in 8 states.

Thompson: Gannett says 28 events in 7 states plus. WaPo says 10 events in 3 states.

Does this new data affect my argument? Yes - a little bit. There really are no major differences between WaPo and Gannett on McCain and Giuliani. And WaPo seems to have a more comprehensive count for Romney than Gannett. So, the effect, if there is one, would only be with Thompson.

The inference I drew about the Thompson campaign is that it had been emphasizing South Carolina, not Iowa, in its public campaign appearances - and that its pivot from the former to the latter is a recent change. The Gannett count would alter this conclusion if and only if: (a) it counts public (i.e. non-fundraising) appearances that WaPo missed AND (b) a comprehensive count of Thompson's public appearances alters the balance between Iowa and South Carolina that the WaPo data indicates. For instance, if those 18 additional Thompson events were public appearances, most of which were in Iowa and not South Carolina - that would change my conclusion.

But, the article from the Tennessean does not give a detailed breakdown of the types of events Thompson held. So, maybe the conclusion would change. Maybe not. At the least, this differing count introduces some variance into my estimate. My intuition at this point is that the differences between the two are due largely to fundraisers - that WaPo's count is basically accurate when it comes to public appearances.

Nevertheless, WaPo is clearly missing some public appearances. This is from my emailer Paul:

Here's an account of him campaigning in SC the weekend of Friday, November 24. No mention on WaPo of Fred doing anything on the 24th, 25th or 26th.

Here's video of Fred speaking at the Florida Family Policy Council on 11/16/2007. Nothing listed on WaPo.

Here's video of him campaigning in Ohio on 12/7/07. WaPo has him as "no events scheduled".

So, WaPo is clearly not counting Thompson's public events comprehensively. If the undercount is not factoring in appearances in Iowa - my conclusion from Friday would not hold. As I do not know what WaPo is missing, I am necessarily less confident about the Thompson campaign.

I certainly would have mentioned Gannett's count if I had known about it. I had assumed that WaPo's count could be considered roughly accurate - and that its numbers for public appearances are reliable. If anybody knows of another count that might enable a verification of the WaPo numbers, please let me know.

Posted by Jay Cost at 01:44 PM | Permalink

A few months back I put together an analysis of Rudy Giuliani's campaign schedule. This enabled me to conclude that, contrary to the conventional wisdom, Giuliani had not recently decided to campaign in New Hampshire. Instead, his strategy had been largely consistent - mixing trips to the Granite State and the states of Super Tuesday.

I would like to update and expand this analysis now. Today, I am going to review all five major Republican candidates. In so doing, I hope to tease out some insights into how each campaign organization views the race. This analysis shall cover candidate visits from November 1 through yesterday - excluding fundraising events and debates. All data was obtained at the Washingtonpost.com.

We will start with Giuliani and Romney. From a travel perspective, they are the most interesting campaigns. Not only do they have the most money, they are also viable in the most states. So, it is most interesting to see how they choose to use their time.

First, we should note that Romney simply campaigns more than Guiliani. Romney has held 78 functions since November 1. He has taken only 12 days off during this period. Giuliani, on the other hand, has held 45 functions since November 1. He has taken 19 days off.

In some respects, Giuliani's campaign schedule has closely matched his poll numbers. He tends to avoid states in which he is losing badly, as well as states in which he is winning by a large margin. So, for instance - he has made only four appearances in Iowa since November 1. His last appearance there was on November 14. Interestingly, he was in Des Moines on December 3 for the NPR debate, but he did not schedule any further events in Iowa. The same goes for this week. He was in-and-out just for the debate.

Relatedly, he has been largely absent from Florida. He has visited the Sunshine State just four times since November 1. Given the importance of Florida to Giuliani's campaign strategy - there are two inferences we might draw from his absence. Perhaps the Giuliani campaign senses that its position in Florida is strong. His lead there is as large as it has been since May. On the other hand, perhaps the Giuliani campaign intuits that Florida is not yet "ripe." Florida voters might not be paying enough attention right now to make campaign stops worth his effort. In this scenario, his time would be better spent in a state like New Hampshire. My guess is that both answers have some validity to them. The Giuliani campaign senses that it has laid a good groundwork in Florida, and so it can save its trips there for later.

But in other respects, Giuliani's campaign itinerary has left me puzzled. In particular, Giuliani has made no stops in Michigan since November 1. He is up in the RCP average, but not by a large amount. I have no answers to this puzzle. I have only questions. Does the Giuliani campaign know something that the poll numbers are not telling us? Perhaps his position is stronger than the polls show, thus enabling him to ignore Michigan. Perhaps it is weaker, thus making trips there not worth his while. Are Michiganders like Floridians - not yet on-line in the same way that voters in Iowa and New Hampshire are? Does the Giuliani campaign simply not regard Michigan as a game changer?

Just as puzzling is Giuliani's de-emphasis of South Carolina, which has merited just two appearances since November 1. South Carolina is legitimately a five-way contest between Romney, Giuliani, McCain, Huckabee, and Thompson. And yet Guiliani does not seem to be campaigning there very hard. All of the questions that we asked about Michigan can be asked once again - with one additional hypothesis. Perhaps the Giuliani campaign thinks that, regardless of what happens in South Carolina, it cannot lose. In a five-man race, it is unlikely that the winner will win by enough to develop any momentum.

So - where has Giuliani been spending his time? Essentially, he has been toggling between New Hampshire, where he has made 18 public appearances since November 1, and the Super Tuesday states, where he has made 11 public appearances. This is consistent with the hypothesis about Giuliani's campaign I made in October. It is striking a balance between the early states and the states of February 5. This strategy has come into better focus now - whereas Giuliani was largely spreading his time between Iowa, New Hampshire, Florida, and South Carolina when I last wrote, it now appears that New Hampshire is the focus of his efforts.

Romney's campaign is less mysterious and much more systematic than the Giuliani campaign. It has done what it claimed all along it would do: attempt to build momentum in advance of Super Tuesday. And so, witness the near-perfect balance in Romney's campaign schedule:

Iowa: 26 appearances
New Hampshire: 22 appearances
Michigan: 5 appearances
Nevada: 4 appearances
South Carolina: 4 appearances
Florida: 12 appearances

This makes perfect sense. There are no anomalies here. The Romney campaign has balanced the early states precisely according to their prestige and their places on the calendar. Accordingly, the bulk of its efforts have been geared toward Iowa and New Hampshire (48 out of 78 appearances have been made there). Meanwhile, the minor states that separate New Hampshire from Florida have received 13 appearances in all. Florida has received 12. Whereas the Giuliani campaign itinerary leaves us with questions - there should be no doubt what the Romney campaign is attempting to do.

What interests me is whether and how either candidate's schedule will change because of the emergence of Mike Huckabee. Most of my data set covers the period before Huckabee opened up a lead in Iowa and took a strong second place in the national polls. It remains to be seen if and how Huckabee will influence the choices of Giuliani and Romney. I wonder in particular if the Romney campaign had hoped to have Iowa "locked up" by this point - and therefore could begin to invest its time in New Hampshire, South Carolina, and Florida. It is, of course, too early to do anything but speculate. I will update this analysis prior to the Iowa Caucuses so that we can get a sense of any late changes in the strategies of the candidates. In the meantime - I would note that Romney has been in Iowa for most of the week.

As I indicated above - I ran a similar analysis on the three GOP candidates whose campaigns are more regionally based: McCain, Thompson, and Huckabee.

McCain's activities are largely unsurprising. His campaign has made 63 public appearances since November 1 (excluding debates). This is more than Giuliani and almost as many as Romney - which is impressive, as McCain is the only major Republican candidate who is not unemployed! By and large, his campaign has emphasized the two states in which it is strongest. McCain has made 32 appearances in New Hampshire, and 15 in South Carolina. Prior to this week, he had not appeared in Iowa since early November - though he did schedule three events around the debate.

There is one quirk with McCain's strategy. I am surprised that the campaign has not spent more time in Michigan, which provided McCain with a big win in 2000. The Democrats, as we all know, are avoiding Michigan. Most of them will not even be on the ballot. Importantly, Michigan is an open primary. Team McCain might be able to count on independents coming to support him as they did eight years ago, given that the Democratic contest probably will not draw their attention as many expect it to in New Hampshire. And yet, it has not held a public event in Michigan for weeks.

Fred Thompson has made only 14 public appearances since November 1 - half of which have been in South Carolina. He has made but two public appearances in Iowa since Halloween. Thus, it seems to me that his decision to throw everything to Iowa is a tad ad hoc. There is no indication from Thompson's schedule that his campaign had taken Iowa so seriously. Is this a recent realization? Did the campaign once think that it could pin its hopes on South Carolina, but has since changed its mind? If it did, it was wise to do so. I do not think it can survive for two weeks and four major contests without placing.

Huckabee has made 44 public appearances since November 1 - fewer than Romney, McCain, or Giuliani. Honestly, this surprised me. I assumed that - as an upstart/insurgent candidate without much money - Huckabee would really be pounding the pavement. Not so much. I was also surprised by the distribution of campaign appearances. I assumed that Iowa would dominate his public itinerary. Again, not so. By my count, Huckabee has made just nine public appearances in the Hawkeye State since Halloween. In contrast, he has made 12 appearances in South Carolina, and 15 appearances in New Hampshire. This is not a huge surprise: Huckabee has visited South Carolina and New Hampshire many times since the beginning of the year. What is surprising is that Iowa has been deemphasized a bit. Whereas for the entire campaign season, Iowa has seen more Huckabee visits than any other state - in the last 45 days, it has received the third most.

So, what inferences can we draw from this? By and large - it serves as confirmation of our previous suspicions. But there are a few new insights to be had. I think there is no doubt that the Giuliani campaign has abandoned Iowa. I also think there is no doubt that the Thompson campaign's decision to embrace Iowa was a recent decision, largely inconsistent with its previous trips. We'll have to watch Romney's schedule this month - and see whether it is narrower because of the threat of Huckabee. Conversely, it will be interesting to see if Huckabee continues to emphasize New Hampshire and South Carolina. Does he think he can get a big boost in the Granite State from an Iowa win? Perhaps he does. Finally - I wonder about the choices of McCain and Giuliani. Hardly any trips to Michigan from either of them. That seems quite strange to me. What are those two up to? One would think that Romney's troubles in Iowa make vigorous campaigns in Michigan seem more worthwhile, especially with those two - but so far, very little attention has been paid there.

Posted by Jay Cost at 12:00 AM | Permalink

Yesterday's Republican primary debate was awful.

There was basically no "hitting" in it whatsoever. Now, this is not necessarily a bad thing. In fact, if the only thing you knew about this debate was that there were no real shots taken by any candidate - you might guess that it was a worthwhile ninety minutes that educated voters on policy differences between the candidates.

But you would be wrong.

In fact, this debate combined the worst feature of the prior outings - questions that elucidated little in terms of substance - with none of the entertaining fisticuffs that at least made the previous debates fun to watch.

There were two major problems.

First, the questions stunk. Just plain stunk. Here are the questions I had jotted down in my notes:

"Does our financial situation create a security risk?"

"What sacrifices should Americans make to reduce the national debt?"

"Who is paying more than their fair of taxes?"

"How would you have an open White House?"

"Ron Paul, how would you get a Congress that completely disagrees with you to go along with your policies if you were elected President, which will never happen, anyway."

"Is it more important to be a social conservative or a fiscal conservative?"

"Alan Keyes, please tell us what you think because your opinion is worth hearing."

What a waste. What were these questions designed to do? They might have some value in a debate between a Republican and a Democrat - for instance, it might be interesting to hear differences between the parties on the relationship between economic and national security. But when you have candidates from just one party participating, you have to try a little bit harder to get them to differentiate themselves from one another. For how pompous the moderator seemed - shushing candidates left and right, and abjectly refusing to allow Fred Thompson to speak on global warming - you would think she was asking something better than these inane queries.

When the questions were not completely useless - the format impeded anything approaching an intelligent answer. The Des Moines Register took the same basic MSNBC format - where candidates are awarded for pithy one-offs and silly sound bite attacks - but did not ask the questions that facilitate those small-ball answers. This was the second big problem. The format. The Register wanted important answers compacted into the petty time allowances. That just was never going to happen. So, Mike Huckabee was given ten or so seconds to tell us something new about how his faith would inform not just his policies generally, but his health care and his education policies.

The last question was the best example of what seemed to me to have been an ill-conceived debate. It was basically an open opportunity to take a shot at another candidate. However, the question itself ("Name a New Year's resolution for one of your opponents") was so poorly constructed that virtually no candidate made use of it. Tancredo, no stickler for subtlety, was the only one willing to take his shot - but the moderator cut him off!

It is not surprising to me that Thompson was given a gold star for his performance. His campaign is, as I have argued, a campaign against the way the media presents politics to the voters. So, it figures that Thompson was quick to complain about the format - and win kudos for expressing what all of us were feeling at that moment.

All in all, the number of good debates have been pathetically few. The press needs to reevaluate its role in our presidential politics. At least so far this cycle, it is not performing the role we might expect of an institution protected by the Bill of Rights.

As for winners and losers - John McCain and Rudy Giuliani. They lost. Big time. The reason is that neither of them is actively competing in Iowa anymore, so they made special trips to the Hawkeye State. They lost a full day of campaigning for this waste of time. With just a month left and both of them in second place - that's a real loss. The winner was clearly Mike Huckabee. This debate could not have changed any minds in Iowa - as he has the lead right now, he benefits the most.

Posted by Jay Cost at 12:00 AM | Permalink

Tomorrow I am going to offer another analysis of "who hit whom." I did this for the last GOP debate, and I thought it gave us some purchase on how the candidates view the race.

Today, I want to lay down some predictions on how the attacks are going to come. My basic intuition is that - whereas in the previous debate, Giuliani received most of the hits - he will probably not be attacked as much today. Instead, look for Romney and Thompson to go after Huckabee. Both of them rightly perceive the former Arkansas governor as a threat.

Generally - what we should see today is part of the final act of this cycle's Republican drama: who shall be the social conservative alternative to Giuliani? This is essentially a battle between Romney, Huckabee, and Thompson. For most of the campaign season, it appeared as though Romney was the presumptive anti-Rudy. At least as far as the polls go, this position is now a toss-up between Huckabee and Romney - with Thompson a marginal player. My guess is that most of the attacks will come from Romney and Thompson - and they will be directed at Huckabee.

I will be particularly interested in two elements of this interaction. First, will Romney's attacks on Huckabee be artful? As I have argued before, I think the Romney campaign has a better grasp of the science of politics than it does the art. Today, I think Romney's hits will have to be dexterous to be effective with Iowan viewers. We know Romney can hit Huckabee. Can he hit him well? We might not get an immediate answer to this question because the Washington insiders who report and analyze these debates tend to relish the hits more than the voters in flyover country. So, we might see something akin to the first Bush v. Gore debate: what might initially be scored as a success by the pundits turns out to be a failure.

Second, how will Huckabee handle the religion issue? The question about Jesus and Satan that he posed to Zev Chafets of the New York TImes Magazine might have been innocent and accidental - but it could just as easily have been purposeful. This leaves me wondering whether and, if so, how Huckabee intends to use religion to marginalize Romney. My guess is that he lays off the religion issue - but that he does so in a way that is a de facto shot. For instance, if asked if Mormonism is part of Christianity, I would expect Huckabee to say something like, "I don't know." Generally, look for Huckabee's attacks on Romney - if there are any at all - to be passive aggressive.

Ultimately, with Huckabee, Romney, and Thompson caught up with one another - McCain will likely go unassailed today - as was the case last time. But this social conservative squabble also means that Giuliani could escape most of the hits, too. He is not really a target right now.

Accordingly, this debate should be very liberating for both Giuliani and McCain. I expect them both to do well. Not only will they likely have nobody attack them - they have fewer audience constraints, too. Romney, Huckabee, and Thompson are actually playing to two audiences. On the one hand, they have to court the undecided Iowa Republicans tuning in (and my guess is that this debate will grab their attention). On the other hand, they also have to court the Washington journalists/pundits who are going to summarize the debate for the rest of the nation. Giuliani and McCain are really just playing to the latter group - as they stand no chance in Iowa.

Posted by Jay Cost at 11:12 AM | Permalink

Hillary Clinton's chief pollster, Mark Penn, offered another strategy memo this week. This time, he discussed the polls - and suggested a way to interpret them. Today, I would like to respond to this memo.

Let's take it point-for-point.

What's happening in the Democratic primary for president?

A lot less than the headlines would suggest.

Iowa continues to be a competitive race while Hillary is maintaining meaningful leads in all the other states and in the national polls that are representative of her Feb 5th strength.

But with the plethora of polls it is becoming increasingly difficult to follow what is a trend, what is a poll without a trend, what is a screened phone poll and what is a computer driven poll. The natural tendency is for those polls that show it closer to get more attention. They are "news."

The big paragraph at the end was what got my attention, and actually induced me to pen a response. Here, Penn mixes a controversial statement with a non-controversial statement - suggesting that the latter implies the former.

I think we would all agree that the media focuses on the polls that show a closer race. However, the first sentence of the paragraph - when we combine it with the penultimate paragraph - is actually quite controversial. It seems to me that Penn is suggesting that if we select our polls judiciously, rather than doing what the headline-hungry press does, we will see that Clinton has a "meaningful" lead in "all the other states" beyond Iowa.

Ironically, the word "meaningful" is, at least for the moment, a meaningless term. Does he mean that her lead is statistically significant? Does he mean that the lead is not just significant, but durable? We cannot yet tell.

Let's see where he wants to take this.

There's yet a new case-in-point of poll confusion today with the release of a slew of Mason-Dixon polls - but a look at their past polls paints a very different story than at first glance. For example, they have Hillary ahead by 3 points today in SC and pundits suggest that this shows how the race has closed. But while other polls showed a strong lead in June, the Mason-Dixon poll had Hillary losing by 9 points in June, so this actually shows Hillary's margin up by 12 points from their last poll and surging. When you look at the facts by tracking results over time from the same poll, she is up, not down. Other polls give her a much wider lead than Mason-Dixon: the latest Pew poll has Hillary ahead by 14 points in South Carolina and the latest ARG poll has her 24 points ahead.

Because predicting primaries is extremely difficult and everyone has their own methodologies, you have to look at polls from the same pollsters to see if there have been changes.

Similarly, the Mason Dixon poll in NH shows a close race with a 3 point lead for Hillary - but their last poll in June gave her a 5 point lead - and a WMUR/CNN poll around the same time had Hillary leading in New Hampshire by 15 points. So Mason-Dixon was low in June and they actually show no statistically significant change in her margin now.

Again, Penn mixes controversial with uncontroversial statements - hoping that the latter validate the former. Here, he suggests that we examine the trend in a given poll to get a sense of the race. This is a valid idea - and it is why Tom, Blake, and Reid are always careful to mention trends when they report new polls on the RCP Blog. However, if we do what he suggests - "look at polls from the same pollsters to see if there have been changes" - we will not find what he asserts - "she is up, not down." At least not if we do it in the correct way.

Penn wants to select June as his baseline. This is not a viable selection - for two reasons. First, our interest is, as he claims, to see if there have been "changes." If we want to identify changes, we should actually look at the polls that immediately preceded the most recent batch. Second, the race was much tighter in June than it has been up until recently. The RCP average of national polls showed Clinton with a lead of less than 10% on Obama as of June 10. I think we would all agree that Team Obama would love it if the race right now resembled the race in June!

So - what we need to do is compare the latest polls to their immediate predecessors. Let's do New Hampshire first.

Zogby had Clinton up by 15 in September. Now it has her up by 11.

Marist had Clinton up by 22 in November and 22 in October. Now it has her up by 13.

Rasmussen had Clinton up by 10 in November and 16 on October 23. Now it has her up by 7.

ARG had Clinton up by 18 in October and 19 in September. It had her up by 11 in November.

CNN had Clinton up by 23 in September. It had her up by 14 in November.

What about South Carolina? It is more difficult to track trends there because fewer polls have been taken - and polling companies are surveying with less frequency. Nevertheless, there are three companies that have polled multiple times in the past few months: Insider Advantage, Rasmussen, and ARG. Insider Advantage and Rasmussen show Clinton's lead shrinking. ARG shows it increasing.

Penn only mentions the ARG poll.

Let's continue with the memo.

2nd case in point: Last week three polls of New Hampshire Democratic primary voters were released. One showed Hillary with a 14 point lead (Marist), one showed Hillary with an 11 point lead (Zogby) and one gave Hillary a 6 point lead (ABC/Washington Post). Which poll got the most attention? The one that showed the closest race - ABC/Washington Post. And poll junkies should also note that New Hampshire polling is particularly difficult because it is often unclear to the last minute which independent voters are coming to which primary - and Hillary has a strong and energized lead with Democrats.

I think Penn gives us a clue here by what he means with the word "meaningful." We learned that Hillary has a "meaningful" lead. Now we learn that Hillary has a "strong and energized" lead. It seems to me that he is not hinting at simple statistical significance. He is hinting at durability. That is - Clinton's lead now can be expected to endure to Election Day.

I would ask how he could possibly know that from the methodology he offers. The only suggestion that he has made so far is that we watch the changes in a given poll. But we will only wind up with the conclusion of meaningful, strong, and energized if we look at certain polls (ARG) and ignore others (Mason-Dixon, Rasmussen).

How do we evaluate his claim?

Penn noted above that each poll has its own particular methodology. This makes it difficult to evaluate Penn's argument - or any assertion about the state of the race. Pollsters are not like academics, who often spend pages and pages describing and defending their methodologies before they actually get to their results. Polling methodologies are treated as proprietary interests, and kept close to the vest. ARG, for instance, gives us almost no hint of how it goes about conducting its polls. Thus, all that we have are these bare numbers, with no indication of how they were created.

Which poll should we select?

Here's what I do. Let's take New Hampshire as an explicative case.

Suppose that we deem five polls to be current: Mason-Dixon, Zogby, ABC News/WaPo, Marist, and Rasmussen. Let's also suppose that we have no a priori idea which poll's methodology is the correct indicator of the preferences of the January 8 electorate (but we assume that one is indeed reasonably unbiased).

So:

If we guessed that Mason-Dixon has the correct methodology, we would expect Clinton to be at 30%.

If we guessed that Zogby has it, we would expect her to be at 32%.

If we guessed that ABC News/WaPo has it, we would expect her to be at 35%.

If we guessed that Marist has it, swe would expect her to be at 37%.

If we guessed that Rasmussen has it, we would expect her to be at 33%.

We have five numbers, each of which comes from a different methodology. We do not know which comes from the proper method - but we do know that each has a 1 in 5, or 20%, chance of being derived from the right method.

What should we do?

We should work to minimize our expected error. What number could we select that minimizes the likelihood that we have chosen incorrectly? Penn wants you to take the 37% and ignore the 30%. That is not the way to minimize your expected error. One way to minimize your expected error is by calculating the statistical average, which is computed as:

(30 X .20) + (32 X .20) + (35 X .20) + (37 X .20) + (33 X .20) = 33.4

[Note: We would get the same estimate regardless of how many polls we believe have accurate methods - so long as we assume that there is at least one.]

This number also happens to be the current RCP average of Clinton's position.

A-ha! Now we have a way to get not just a sense of the trends within a poll, but we also get a sense of what all the polls are pointing to. It is not necessarily an unbiased estimate. After all, there may be a poll in there with an absolutely lousy methodology. If there is, our estimate is indeed biased. However, our average minimizes this bias in light of the fact that we have no a priori idea which poll (if any) has lousy methodology.

I cannot emphasize enough that this solution is not ideal. It is, rather, the most practical response to the choice of polling companies not to publish their methodologies in any great detail, and therefore to thwart a robust debate among experts as to which is superior. If pollsters had the same reporting protocols as, say, the American Political Science Review, we could improve upon this average by evaluating each poll's methodology - and selecting those whose methodologies are most sound. But, because we are not professional pollsters and cannot go "behind the veil," this is the best we can do.

So, we compute the statistical average. This is the best response to the polling environment. We can compute this average over time, graph it, and get a quick visual idea of where the race stands from this "minimize expected error" perspective.

You probably have seen such graphs before:

So, contrary to Penn's claims - we see that the race has tightened in New Hampshire. If we graph the other contests, we would see that most (but not all) show tightening of various degrees.

And therefore we see that Penn's initial uncontroversial statement about the media's reporting of polls does not validate his controversial statement that Hillary's lead is meaningful, strong, and energized. The latter conclusion comes only when one selects the polls that show Hillary at her strongest. This is not the correct way to get a sense of the race.

Posted by Jay Cost at 12:20 PM | Permalink

Jonathan Martin had an interesting column in Sunday's Politico about the GOP race in Iowa. This is what he had to say:

As the battle for Iowa enters the home stretch, the race appears to be breaking down along a simple fault line: Mike Huckabee's momentum and passion versus Mitt Romney's organization.

What one has the other lacks. [snip]

All the mojo seems to be with Huckabee - a Newsweek poll released Friday shows him ahead of Romney by a staggering 39 percent to 17 percent margin.

But Iowa veterans warn that while Huckabee is all the rage, it's almost impossible to win in Iowa without an organization - and Huckabee's is skeletal. "A poll is a poll in Iowa," observes Ed Failor Jr., a longtime Iowa Republican and anti-tax leader based in Muscatine who is currently undecided. "But it's different than turning out voters on caucus night."

I generally agree with Martin's characterization of the Iowa race. However, I think there are a few important caveats to make.

First, Huckabee is still in what might be called a "honeymoon" phase. Voters are just now learning about him. This is important because when voters do not know very much about a candidate, they tend to ascribe to him the qualities that they prefer - even if the candidate might not possess those qualities. It is kind of like projection. Larry Bartels found this to be the case for Carter in 1976, Bush in 1980, and Hart in 1984. My intuition is that something like this is happening with Huckabee. If so, that would be good news for the Romney campaign - which can use its resources and its remaining time to educate Iowans about how Huckabee is not the best thing since sliced bread. And so, Huckabee's numbers might be artificially high right now.

Second, I remain unconvinced that Huckabee will necessarily need a strong GOTV organization to win the caucus. Last month's ABC News/WaPo poll found that Huckabee has a huge lead among those who have previously attended a caucus. Romney, on the other hand, was winning among those who have never previously attended a caucus. So, Huckabee might have found a core of supporters who do not need to be mobilized actively, while Romney might need to use his full GOTV efforts just to get to the numbers he's registering in the opinion polls. Relatedly, this caucus is probably the most intense in the history on the Republican side. High intensity campaigns can have the effect of increasing voter turnout. We might find, in other words, the same kind of irony that manifested itself at Ames, where Romney and Brownback worked to get people to the straw poll, some of whom supported Huckabee.

Third, I think it is wrong to discount the passion of Romney supporters. I'd wager that Romney has a floor of something around 20% in these polls - and that is an important asset in a five-way race.

Fourth, I will be interested to see what happens with Fred Thompson - who apparently has decided to make a final stand in the Hawkeye State. As Stephen Hayes wrote on Friday - Thompson has decided to spend every day but Christmas campaigning in Iowa. He does not believe that he can wait until South Carolina - whose primary is sixteen days after the Iowa Caucus - to have a decent show. This is probably a necessary move for Thompson, who will probably not win Iowa - but, this effort could make it so that he places respectably. Hayes writes:

Thompson has said publicly that he needs to finish in the top three in Iowa. Campaign officials say that a strong third place finish--presumably behind new frontrunner Mike Huckabee and former frontrunner Mitt Romney--would likely give them enough momentum to survive New Hampshire and compete in South Carolina and beyond. A second place finish would be a victory. "Just when the interest is there the greatest, is when we'll be here the most."

Finally, the Huckabee campaign gets an A-for-effort for this little gem, but I am not buying:

It's enough to delight Huckabee's camp, but also cause to make supporters worry that expectations are rising too high too soon.

"Folks need to calm down here," says Eric Woolson, Huckabee's Iowa director.

"We've said from day one, back there in January, the objective is to finish in the top three. It's always been to finish in the top three and that's still the objective today."

Nice try. I always get a kick out of campaigns trying to manage expectations - especially when they make silly statements like this (which they often do). If Huckabee finishes in third, he is finished.

Posted by Jay Cost at 12:00 AM | Permalink

There are two ways to evaluate Romney's speech. The first is as a matter of political theory: is the logic of the speech sound? The second is as a matter of public opinion: will its logic have any force with the voters?

As a matter of public opinion, a few voters who watched the speech were surely swayed - but it will not have a direct, mass effect. Not enough people will have watched it - even in Iowa. What it might do is change the media narrative. Romney can't have people in Iowa and New Hampshire continuously hear that he is slipping in the polls. This speech can change that story, at least for a while.

As a matter of political theory, I have mixed feelings about it. Romney's basic thrust was the following. There is a common faith in this country, a "great moral inheritance" derived from the shared belief in a divine, benevolent Creator, that translates into universal political ideals: equality, service, and liberty. The American position therefore recognizes that faith is an important aspect of civil society, but that at the same time society must allow for the multiplicity of religious sentiments. This creed is what has fostered a dynamism that "has kept America in the forefront of civilized nations even as others regard religious freedom as something to be destroyed." Accordingly, what is needed is a president who appreciates and shares this basic faith. Furthermore, citizens have a duty to recognize both the importance and the limitations of the role of faith in public life: candidates should be judged on the basis of the fact that they have faith, but not on its particulars.

I think Romney has hit upon one of the original premises of American religious toleration. I thought he did the Founders' views justice - and he also made them relevant to today. On the question of how to integrate religion into the basic structure of civic life, I think Romney's reasoning was sound.

But this is the view from 30,000 feet. The purpose of this election is not to design a new constitutional system. Its purpose is to elect a president to govern over a divided nation. Romney offered a rigorous defense of the foundation of American civil society - but he never addressed the concern that induced him to give this speech in the first place.

And what is that concern? It is the same concern that always turns American unity into partisan division - the transition from questions about how to structure the government to questions about what to do once the government has been structured. Here - we are confronted with divisions, many of which have their derivations in differing religious opinions. While it is true that religious similarities yield similar political ideals - it is also true that religious differences yield different political preferences.

And herein lies Romney's essential problem. He has taken issue positions that many voters take because of specific religious beliefs. This is not to say that there are not other ways to derive those positions - but it is to say that many people who adopt those positions justify them by their particular theological beliefs. They are not, cannot, be justified by a shared religious creed. If they could, everybody who holds to that creed would be in agreement, and there would be no political issue. You can justify trial by jury or "guilty until proven innocent" by reference to this shared American religious creed. But you cannot justify opposit