Debates do matter — sometimes

Our yen for “yes” or “no” answers sometimes clouds our understanding — in this case, of presidential debates. Observers seem wedded to saying either “Yes, debates do have an impact” or “No, they don’t.” Reality is more complicated. 

A Gallup review “reveal[s] few instances in which the debates may have had a substantive impact on election outcomes — public preferences moved quite a bit around the debates, but the debates did not appear to alter the likely outcome.” Of course, one reason debates have had minimal impact on outcomes is that few presidential races are close. Over half the elections since 1960 have been decided by margins of over 7 points. Altering those outcomes would have required debates to create huge swings. Only five elections were decided by margins of less than 4 points.

Those 4 points are relevant because, using Gallup’s data, that’s the average change in the margin pre- and post-debate. Of course, Gallup’s aren’t the only polls conducted, and we know those polls often provide divergent data. A more comprehensive analysis by Professors Robert Erikson and Christopher Wlezien found that “Debates seem to make at least a bit of difference,” though “any claims about debate effects … are fragile.” (Interestingly, a number of those searching for certainty in arguing for no debate impacts have quoted the second part of the professors’ statement, but failed to take note of the first.)

The Bush-Gore debates of 2000 are one of the two sets Gallup analysts regard as influential (1960, where data is much thinner, is the other). In a panel study that reinterviewed the same individuals before and after the debates, Professors Sunshine Hillygus and Simon Jackman detected a change in the margin of about 4 points in Bush’s direction. But which voters were most affected? Just 9 percent of Bush voters and 13 percent of Gore voters moved either to undecided or to their original candidate’s opponent. Those were mostly mismatched partisans — i.e., Republicans who had been for Gore, and Democrats who had favored Bush — a segment that barely exists in today’s political world. Those undecided going into the debates split evenly — 21 percent moving to Bush and 21 percent to Gore. 

The story of the 2000 debates is more tortured in Gallup’s data. In its telling, the debates produced considerable short-term movement, the effects of which faded quickly. Gore went into the debates leading in the popular vote and ended up winning the popular vote quite narrowly. Of course, had Gore maintained the 8-point lead he enjoyed going into the debates or even the 5-point lead he held shortly after the first encounter, he would have undoubtedly won the Electoral College and the White House. 

So it’s just possible that the second and third debates moved enough voters to hand Bush the presidency, but there is no way of knowing for certain whether those gyrations in the numbers were produced solely or even mainly by the debates — or even how real they were at all. The ABC News poll suggested almost no movement after the second and third debates, with Bush holding the lead before and after both. Pew reported that Gore moved from a 1-point deficit to a tie after the third debate, a margin-of-error difference, to be sure, but one that suggests a different interpretation of events.

In short, there are no clear answers to how much debates matter. We can be pretty certain they haven’t produced 8- or 10-point shifts, but there is evidence that in an otherwise close race they can move enough voters to make a difference. 

Trying to predict the future is fun, but let’s just wait and see what actually happens before claiming certainty.

Mellman is president of The Mellman Group and has worked for Democratic candidates and causes since 1982. Current clients include the majority leader of the Senate and the Democratic whip in the House.