If you were to rank the most exciting presidential nomination contests, the 1996 Republican race would be near the bottom. Bob Dole, the “next-in-line” GOP candidate and the Senate majority leader, built up a huge lead in polls and endorsements early in the race and was never seriously challenged for the nomination. Dole did lose the New Hampshire primary by a single point to Pat Buchanan. But the field soon consolidated around him, and he went on to win 44 of 50 states.
And yet, contemporaneous accounts of the sleepy-seeming 1996 campaign portrayed it as incredibly dramatic, full of “unexpected” twists and “unpredictable” turns. Take this March 7, 1996, article from The New York Times, for example — it was written after Dole had won 12 consecutive primaries and caucuses in the previous week. There are four expressions of surprise in a single paragraph: It’s taken as shocking that the early primaries were as competitive as they were, but equally “striking” that Dole rebounded so quickly from them.
After an unpredictable early stretch of primaries, where candidates seemed to flicker out like trick birthday candles, only to re-ignite unexpectedly, Mr. Dole’s return to a commanding lead so early in the voting was striking. The positioning and sorting of the field yesterday was particularly unusual: Almost simultaneously, Mr. Lugar and Mr. Alexander bowed out, as Governor Bush and Mr. Kemp put forth their dueling endorsements.
I don’t mean to pick on this article, which happens to have been written by a terrific journalist,1 but it’s typical of the breathless fashion in which developments on the campaign trail are reported. There is a constant series of overcorrections in the conventional wisdom. In this case, because the initial threat to Dole was overstated by the press — Buchanan, a factional candidate, had little chance to see his support grow beyond the 27 percent of the vote he won in New Hampshire2 — Dole’s “comeback” was incorrectly portrayed as unexpected and dramatic.
These biases hold in coverage of the general election too, of course: There were 68 purported “game changers” in the 2012 general election, most of which turned out to be irrelevant. But for the political observer trying to sift faux game changers from genuine twists in the campaign, the primaries present a couple of additional complications.
First, there are far more opportunities to be “surprised” in the primaries. Let’s start with the most basic stuff. In a nomination race, there might be a dozen or more candidates, instead of just two. And states vote one at a time, instead of all at once.
Furthermore, in a nomination race, there is an abundance of metrics by which you might judge the campaigns: national polls, Iowa polls, New Hampshire polls, favorability ratings, endorsements, fundraising, staffing, even crowd sizes and yard signs. Eventually, we’ll also be able to look at delegates, which can be counted in many different ways. For any of these metrics, you can report on the level of support (“Hillary Clinton is polling at 48 percent”), the trend (“she’s lost 4 percentage points since last month”), or even the second derivative (“she’s losing ground, but not as quickly as before”). Multiply 23 candidates3 by 10 metrics by three ways of reporting on those metrics, and you have 690 opportunities to be “surprised” at any given time.
In a sense, the primaries are a lot like the NCAA basketball tournament: You know there are going to be some surprises. The odds of every favorite winning every game in the NCAA tournament are longer than a billion-to-one against. And yet, in the end, one of the front-runners usually wins. (Since the men’s tournament expanded to 64 teams in 1985, all but three champions have been No. 4 seeds or better.)
So be wary of grand pronouncements about What It All Means based on a handful of “surprising” developments. Is Scott Walker’s campaign off to an “unexpectedly” bad start, for instance? Maybe. (I wouldn’t be thrilled if I were one of Walker’s strategists. I’d also remind myself that we have five months to go before the Iowa caucuses.) Even if you grant that Walker is having some problems, however, it would be stunning if all the Democratic and Republican campaigns were doing exactly as well as pundits anticipated. At any given moment, some campaigns are bound to be struggling to meet expectations, or exceeding them.
Similarly, while one might not have predicted that Bernie Sanders would be the one to do it, it was reasonably likely that some rival would emerge to Hillary Clinton. It’s happened for every non-incumbent front-runner in the past: Buchanan for Dole; Bill Bradley for Al Gore.
The other big difference between the general election and primaries is that polls are not very reliable in the primaries. They improve as you get closer to the election, although only up to a point. But they have little meaning now, five months before the first states vote.
It’s not only that the polls have a poor predictive track record — at this point in the past four competitive races, the leaders in national polls were Joe Lieberman, Rudy Giuliani, Hillary Clinton and Rick Perry, none of whom won the nomination — but also that they don’t have a lot of intrinsic meaning. At this point, the polls you see reported on are surveying broad groups of Republican- or Democratic-leaning adults who are relatively unlikely to actually vote in the primaries and caucuses and who haven’t been paying all that much attention to the campaigns. The ones who eventually do vote will have been subjected to hundreds of millions of dollars’ worth of advertising, had their door knocked on several times, and seen a half-dozen more debates. The ballots they see may not resemble the one the pollsters are testing since it’s likely that (at least on the GOP side) several of the candidates will have dropped out by the time their state votes.
Some reporters object to this by saying that the polls are meaningful to the extent that they influence the behavior of the campaigns: If Joe Biden enters the race because he reads the polls as indicating that Clinton is vulnerable, that could matter, for instance.
Fair enough. But it’s dubious to compare hypothesized behavior with actual outcomes we’ve seen in past election years. If Biden or Mitt Romney thinks the field is weak and makes a late entry, that will be interesting. If Donald Trump or Sanders actually wins the nomination or comes very close to doing so, that will be a watershed in American political history. But there’s a high rate of false alarms compared with the number of late-entry candidates that actually make a bid or watershed moments that actually occur.
None of this is to imply that nominations are all that easy to forecast. And some things this year have been genuinely surprising. In particular, that there are 17 Republican candidates, including a dozen or so who have traditional credentials for the White House, is unprecedented. If you want to develop a theory about how “this time is different,” figuring out how to explain the size of the Republican field (and how it might affect the race) seems like a good starting point.
Oddly, this abundance of candidates has been somewhat taken for granted in campaign coverage, even though it potentially plays a big role in explaining Trump’s position in the polls, among other things. The reason may be that it’s something we’ve known about for a long time; there aren’t a lot of clicks to be had from the headline “17 Candidates Still Running; Nothing Else Much Changed Today.” Chasing down the bright, shiny object is more exciting, but usually not more revealing.