Late Thursday night, we will have the results of the Iowa caucuses (which will not be exactly what they appear to be, especially on the Dem side, see below).
The punditocracy will opine like crazy about who’s damaged and how badly, who has to drop out, how the winners did it and why the others failed.
Advice: Bear in mind that four years after Howard Dean’s sensational meltdown in the last few weeks of the Iowa campaign, activists, pundits and political scientists still have different theories about why it happened. They won’t really know this time either, especially in the first hours after the results come in. So take the analyses, especially the more confident analyses, with plenty of salt.
Even now, without any results to use as tea leaves, some soothsayers are brave and foolish enough to tell you what is going to happen. (I heard David Brooks say Friday on “The Newshour with Jim Lehrer” that if Hillary Rodham Clinton wins Iowa, the Democratic race will be over. Add salt.) This piece does not tell the future (I know the future, I just ain’t telling), but offers a few facts, thoughts, quirks, arguments and advice that, if you don’t already know it all, may prepare you to understand the results when they come in, and to resist the hype.
In a sense, the pundits will be making a big deal about not very much. Very few national convention delegates are at stake Thursday night (and none will actually be chosen in a binding way until June, see below). The difference between first, second and third in delegates is likely to be puny, especially if you think about the number of caucus-goers separating first from third. Historically, fewer than 10 percent of eligible Iowans caucus. In all likelihood, Thursday turnout might approach or exceed 200,000, which would set an Iowa record. But caucus attendees will still comprise much less than 1 tenth of one percent of the U.S. population. The Iowa result doesn’t have to have any more impact on the other 49 states than the media and the non-Iowa electorate choose to give it.
The momentum conundrum
But that’s sort of the point. The media and the rest of the country do give it impact. According to conventional wisdom, it’s the momentum Iowa gives to the winner heading into the rest of the race for the nomination. In 2004, John Kerry came from behind to win Iowa, then New Hampshire, and never faced a serious bump on the road to the nomination after that. Likewise Jimmy Carter, who really put the Iowa caucuses on the map — because his 1976 Iowa win really put him on the map — also went on to win New Hampshire and then build an insurmountable lead. (Weird fact: Carter actually finished second in 1976, to “uncommitted.”)
So you might be surprised to learn that historically it has been rare for the Iowa winner to go on to win in New Hampshire.
In 1980, George Bush the elder declared that his impressive Iowa win (by 2 percentage points over front-runner Ronald Reagan) gave him “the big mo” heading to New Hampshire. But Reagan crushed Bush by 50-23 in New Hampshire and went on to win 28 of the remaining 33 primaries to lock up the nomination. (By the way, that year, there was more than a month between the Iowa caucuses and the New Hampshire primary.) That left time for a second drawn-out campaign of retail politics in New Hampshire. In 1984, the gap shrunk to just eight days, and it remained eight days until this year, when the gap will be a record low of five days. Jan. 3 will also be the earliest date ever for the Iowa caucuses, by a factor of 16 days.
Another case where the “Iowa bump” turned out to be a molehill was the Dick Gephardt campaign for the 1988 Democratic nomination. U.S. Rep. Gephardt built his strategy around winning Iowa, and he did. But the media was more impressed by an unexpectedly strong third place showing in Iowa by former Massachusetts Gov. Michael Dukakis. It was declared unexpected because Dukakis was from Massachusetts, while Gephardt (of Missouri) and Iowa runner-up Sen. Paul Simon (of Illinois) were from states that bordered on Iowa. Gephardt’s Iowa bounce strategy was such a dud that his campaign manager, Bill Carrick, later remarked “it would have been nice if someone had told us that Iowa’s going to be worth Idaho this time around.”
Lest you think I’ve made too much out of a couple of exceptions, what happened to Bush and Gephardt has been very common. Since 1972, when the Iowa caucuses first became an influential event (note that date for the next time someone acts as if Iowa has always had its special powers; it’s not that long a history), there have been 11 contested nomination battles. (That doesn’t count the cases in which an incumbent president was seeking renomination.) In only three of those 11 cases has the Iowa winner won New Hampshire. In those 11 contests, only one candidate, Bill Clinton in 1992, has won a major party nomination after losing both Iowa and New Hampshire. (Clinton finished third in Iowa — fourth if you count “uncommitted.”)
One more key takeaway about how the momentum game is played: The other big factor besides what place a candidate finishes is whether he or she exceeded the last round of expectations (also assigned by the media and the political class). All of the campaigns are currently trying to lower expectations so they can claim Thursday night to have exceeded them. Hillary Clinton’s interview Sunday with ABC’s George Stephanopoulos included this typical exchange:
CLINTON: I think, because it’s so close — you know, when I started here, I was in single digits. I mean, nobody expected me to be doing as well as I’m doing in Iowa. I was running against one opponent who has been campaigning here for four years [that would be Edwards], another opponent from a neighboring state [that would be Obama, of Illinois]. So I believe that this campaign will be bunched up. I think that the history out of Iowa is that a lot of people live to fight another day.
STEPHANOPOULOS: So you may not win?
CLINTON: I’m not expecting anything.
The three-tickets-out oversimplification
Analyzing and prognosticating Iowa has become a field of its own. A few cycles back, pundits adopted a rule of thumb, that there are three tickets out of Iowa (meaning the top three finishers can proceed to New Hampshire without appearing ridiculous). David Yepsen, the dean of Iowa political reporters, has added a coda that of the three ticket bearers, the winner travels first-class, the runner-up gets a coach ticket and the third place finisher flies standby. Yepsen goes on to say that this year, for the Dems, there will be no standby ticket.
Huh? Remember, salt.
On the Repub side, the three-tickets thumb-rule is in big trouble. Unless the recent polls have been taken in the wrong state, either Mike Huckabee or Mitt Romney will win among Iowa Republicans. Whoever finishes third will claim a big victory. There’s a lot of current buzz around John McCain as a surprise third-place finisher. But McCain is certainly going to New Hampshire either way. Rudy Giuliani still leads in most national Republican polls and won’t be dropping out of the race this week or next. So that’s four post-Iowa candidacies. And it seems at the moment that Ron Paul and Fred Thompson will continue in the race.
Perils of polling
In 2004, late polling picked up the fact that Kerry and Edwards were surging while Dean and Gephardt were fading. But only one poll, published by the Des Moines Register on the day before the caucus, showed the top three finishers in the correct order: Kerry, Edwards, Dean.
The Register published Tuesday its big pre-caucus poll. On the Dem side, among likely caucus-goers, it showed Obama: 32 percent; Clinton: 25; Edwards: 24, and all other Democrats in single digits.
On the Repub side, the Register has the race Huckabee: 32 percent; Romney: 26; McCain:13; Thompson: 9; Paul: 9, and Rudy Giuliani: 5.
The Register has a reputation for polling the Iowa caucuses about as well as it can be done. But if that gives you confidence in those results as a likely outcome, you need to read the rest of this section. Reading polls as predictions is always hazardous, and for several reasons that you may not be thinking about (details below), it’s especially hazardous in the Iowa caucuses.
The first problem: As difficult as it is to identify likely voters (who can vote with a relative few minutes of effort in many cases) in a poll sample, it is harder to identify likely caucus-goers (who must spend an entire evening doing their civic duty). Various polls have “found” that, as a percentage of all the adult Iowans they interviewed, the percentage that could be classified as likely caucus participants ranged from under 10 percent in many polls, to more than 30 percent in some.
The second big challenge for those who try to read polls as predictions is a strange, complicated rule on the Democratic side.
The Repub side is fairly straightforward. Soon after the caucus convenes at 7 p.m. Thursday, Republican precincts take a straw poll of presidential preferences and those numbers are reported to the waiting media as the statewide result. Straightforward, simple, but also slightly misleading, because it is nothing but a straw poll. It doesn’t translate into national nominating delegates. The delegates are chosen by a multi-stage process, like Minnesota’s, where caucus attendees choose delegates to state conventions, those conventions choose congressional district delegates, then state convention delegates. The Iowa State Convention in June makes the final choices, presumably long after the identity of the Republican presidential nominee is known. But at least the Republican straw poll resembles an election and the polling is slightly less complicated.
On the Dem side, you really can’t speculate intelligently without understanding the rules of viability and the second round. The rules can vary a bit, but here’s how it worked at the caucuses I attended in 2004 (and the rules are still roughly the same): As the Dem caucus-goers enter, they indicate their presidential preferences. These results are tabulated and mailed to the Democratic state committee, but these results are not reported to the media, then or ever. (Gil Cranberg, former editorial page editor of the Des Moines Register, is very upset about this and railed against it recently in a New York Times op-ed.)
After a few other items of business, the caucus-goers get out of their seats and gather in bunches around the room according to which presidential candidate they support. A first count is taken. In most precincts, any candidate whose gaggle of supporters represents less than 15 percent of the total is declared non-viable, which means that candidate won’t be getting any delegates out of that precinct. That candidate’s supporters are given a chance to join up with one of the viable candidates’ gaggle (presumably reflecting their second preference). The precinct captain for each of the viable candidates gives a speech, imploring the non-viables to come on over.
According to all recent polls, only Clinton, Obama and Edwards are attracting more than 15 percent support among Iowa Democrats statewide. The rule applies on a precinct-by-precinct basis. So Bill Richardson or Joe Biden may have pockets of strength where they will reach 15 percent. But in a given precinct, there could be 10 percent or more of caucus-goers who will make a second choice. In some cases, deals are struck. In 2004, Dennis Kucinich agreed to ask his supporters to switch to Edwards on the second round in precincts where Kucinich was declared non-viable. This may be one reason that Edwards ended up a higher share of delegates (32 percent) that the entrance polls conducted by the media suggested he received of caucus attendees (26 percent).
A few recent polls have asked likely caucus-goers about their second choices, and the results tend to favor Edwards. On the other hand, there is some buzz around the idea that Richardson is trying to position himself to become Clinton’s running mate, and might be motivated to recommend that his supporters move to the Clinton gaggles in precincts where Richardson is non-viable.
And this just in: Although polls suggest that Kucinich doesn’t has much support in Iowa, he issued a press release Tuesday urging his supporters to caucus for Obama on the second round, if Kucinich is ruled non-viable. Of course, there’s no guarantee that anyone will go where they’re told.
In summary, a poll tells you that on the eve of the caucuses, the top three candidates are each supported by 30 percent of likely caucus-goers. The other 10 percent are scattered among five also-rans. Even if the poll has accurately identified a representative sample of likely caucus-goers, it doesn’t tell you enough unless it also tells you who the supporters of the also-rans will decide to support on realignment round.
What the ‘results’ really measure
The figure that the Iowa Democratic Party reports for each candidate’s showing on caucus night is a percentage. The uninitiated probably assume that this reflects the portion of all caucus-going Iowa Democrats who favored that candidate. Not so. Not so even after you account for the second choice realignment round. Why not? The only thing really known on Thursday night will be how many delegates to Iowa’s 99 county conventions will be pledged to each candidate. The county conventions will send delegates to the congressional district level, which will choose the actual state convention delegates.
The percentage reported to the media is a weird statistic that Iowa Democrats call “State Delegate Equivalents.” State delegate equivalents project how many delegates to the Iowa State Democratic Convention in June that candidate is likely to have. Here’s the next catch. Each county has a preset number of delegates to the next level of conventions. The number does not reflect how many Iowans turned out in that county on caucus night. No, it reflects the voting strength of Democrats in that county over the previous two election cycles. Therefore, all Iowan caucus-goers do not have equal weight in earning their candidate a “state delegate equivalent.”
As John Deeth explained in a recent post in the Iowa Independent web site, in 2004 Fremont County earned a state convention delegate for every 22 caucus-goers, while at the other extreme, Johnson County received one state delegate for every 79 caucus-goers. Rural counties generally get more delegates per caucus-goer.
So, theoretically, if candidate A has 1,000 supporters in a rural county and candidate B has 1,000 in an urban county, the candidates might look tied in a poll, but Candidate A will get a lot more State Delegate Equivalents and will be declared a big winner on caucus night.
These strange rulings are among the reasons political scientist Charles Franklin of pollster.com recently wrote that:
“For a fair test of how good the polling is this year, wait for New Hampshire where pollsters poll, voters vote, and they just count up the results.”
Eric Black, a former reporter for the Star Tribune, writes about national and state politics, foreign affairs and other topics. He can be reached at eblack [at] minnpost [dot] com.