By Rick Valelly
So far as I know there were (besides me) only two other political scientists inside the Wells Fargo Center this past Monday night. The other two were from Georgetown University — and they have been on the campaign trail all of 2016 with their students in tow. I think that was it for the political science profession this week in Philadelphia.
On the other hand, literally thousands of journalists, famous and not so famous, could be seen throughout the first floor hallway and inside the arena. In fact I managed to have my own brush with journalistic fame. I shook hands late Monday afternoon with Chris Hayes, the MSNBC news hour host, and told him that I’m a fan (which I am.) He gripped my hand earnestly and looked me in the eyes. “That means so much to me.” Then he swooped past with a staffer twice his age trailing him. (Hey, I knew that he didn’t mean it. I’m not bitter about that.)
So why do political scientists not attend party nominating conventions — but journalists do? Part of the answer is that political scientists usually need a press credential. Most political scientists in America don’t happen to have a generous local community newspaper like this one, willing to put in the time and effort to secure a press credential, as The Swarthmorean did for me. That was how I happened to be smiling at Chris Hayes and having him look at me coolly with an “uh oh, not another eager fan” wariness in his eyes.
But another part of the answer is that political scientists have long thought that campaigns — and by extension nominating conventions and the fall candidate debates — do not strongly affect the final outcome of the general election. Unless you are an avid student of party organization and candidate selection rules why attend the conventions? You may feel like you are watching history-in-the-making (you are.) But it isn’t the kind of history that will change the result on Election Day. That’s PoliSci 101 about conventions.
Conventions certainly are useful gatherings for party activists. They get a chance to satisfy their passions. Conventions matter, too, for the important task of hammering out the platform. Political scientists have found that by the time presidents leave office they have honored, in one way or another, about 80% of the platform on which they ran. Conventions also offer that part of the public which has not been paying much attention up to that point a chance to brush up on what’s going on. There are several nights during which they can start to learn about the protagonists in the contest, particularly about the one that they have some inkling they will vote for.
But nominating conventions — and the campaigns that happen after them up until Election Day — do not make or break a candidacy. This is where the famous political science finding that “campaigns don’t matter” comes in. Candidates are, after all, generally equally skillful (albeit in different ways), equally liked (or hated), equally well-resourced, equally competitive on the ground, and equally good at putting on a prime-time portrayal of their party’s stances. Most important, they have equally high “floors.” That is, in every presidential election year about 80-85% of partisan identifiers in the electorate already know before the nominating convention that they will vote in the end for their party’s nominee.
Even political parties that look divided during the four days of the convention do not really suffer. Most voters — 9 out of 10 — have partisan identification. The independent voter turns out to be mostly a myth; 3 out of 10 people may say they are independents if asked by a survey — but the follow-up questions will sort them into red or blue. Because voters are partisan they will get over any sense of discontent and “come home” on Election Day. The alternative is seeing the other party win.
If campaigns cancel each other out then what are the variables that affect the outcome? Call them “the fundamentals.” There are four: (1) the underlying balance of partisan identifiers (2) the relative partisan cohesion of group bases (such as whites without college degrees, Latinos, African-Americans, evangelicals, and so on) (3) whether the economy is really and actually good or bad, and (4) whether “time for a change” sentiment is strong enough to affect many voter decisions — which it usually is after a political party has held the White House for two terms in a row. Each election features relatively slight but numerically large shifts in these variables. New voters come into the electorate for the first time and choose identification with one party of the other at registration (perhaps because of group identification or perhaps because of the partisan identification of friends or parents.) Voters temporarily defect from their party by switching their votes (or effectively defect by not showing up.) Usually inactive identifiers become passionate about their candidate or truly repelled by one of the candidates. Some of these movements cancel each other out, but the net effect will be in one direction. This is the margin where elections are won.
But isn’t a more nuanced view of conventions possible, one that retains some causal role for campaigns and conventions possible? What about the famous “bounces” that candidates have after their nominating conventions, or after they seem to win a debate?
I thought a lot about bounces back in 1988 when Michael Dukakis had one of the largest bounces in history, went off to Martha’s Vineyard, and took a long (really long) vacation, figuring that he would get back to work after Labor Day. Every afternoon during my Vineyard vacation I jogged by his house and seriously thought about barging in to grab him by the lapels (“Mike, I’m a Swarthmore College graduate too and I need to talk to you!”) and tell him to get back to campaigning. By Labor Day, Dukakis’ astounding bounce had literally vanished. If he had kept up the campaign couldn’t he have hung onto his bounce and turned it into a “bump” — that is, a 1 or 2 point increment in public support that would stick with him until November?
You’ve probably been thinking about bounces, too. Since his nominating convention Donald Trump has gotten a bounce – even though he had a disastrous convention. Hillary Clinton had a great convention this week. Won’t she get a bounce?
Who is going to come out ahead by a point or two after the dust settles? Or will the polls go back to where they were before the conventions, i.e. with Hillary having a slight lead?
Some percentage of the electorate makes up its mind for the first time during the conventions, yes. Yes, by the time surveys have sorted out what vote intentions are after these two weeks it is possible for one candidate to have an edge.
But here’s the but: that edge is fragile and is likely to crumble as other voters who still have not made up their minds end up making a choice in their minds — and start responding to surveys between now and Election Day. The polls are still not accurate; they are going to become increasingly accurate and as they do the bump, if there is one, will vanish. Maybe Dukakis knew that and figured that he could take a well-deserved rest.
Wait, you’re thinking, if that’s true — that the bump can crumble — then that must mean that the campaigning between now and election day matters! Why else would undecided voters be making up their minds between now and then and thereby eroding whatever bump that one or the other candidate will seem to have by next week?
That’s where PoliSci 102 comes in — the newer political science about the ways in which campaigns do matter after the conventions are over. This year is turning out to be a very suspenseful year from the perspective of that newer political science about kinds of campaign — and the difference that different kinds of campaigns historically have made. Tune in next week for that discussion.