That presidential debates can be “game changers” is a belief almost universally held by political pundits and strategists. Political scientists, however, aren’t so sure. Indeed, scholars who have looked most carefully at the data have found that, when it comes to shifting enough votes to decide the outcome of the election, presidential debates have rarely, if ever, mattered.
The small or nonexistent movement in voters’ preferences is evident when comparing the polls before and after each debate or during the debate season as a whole. Political lore often glosses over or even ignores the polling data. Even those who do pay attention to polls often fail to separate real changes from random blips due to sampling error. A more careful study by political scientist James Stimson finds little evidence of game changers in the presidential campaigns between 1960 and 2000. Stimson writes, “There is no case where we can trace a substantial shift to the debates.” At best, debates provide a “nudge” in very close elections like 1960,1980, or 2000. A even more comprehensive study, by political scientists Robert Erikson and Christopher Wlezien, which includes every publicly available poll from the presidential elections between 1952 and 2008, comes to a similar conclusion: excluding the 1976 election, which saw Carter’s lead drop steadily throughout the fall, “the best prediction from the debates is the initial verdict before the debates.” In other words, in the average election year, you can accurately predict where the race will stand after the debates by knowing the state of the race before the debates. Erikson and Wlezien conclude that evidence of debate effects is “fragile.”
Why are presidential debates so often inconsequential? After all, many voters do pay attention. Debates routinely attract the largest audience of any televised campaign event. And voters do learn new information, according to several academic studies. But this new information is not likely to change many minds. The debates occur late in the campaign, long after the vast majority of voters have arrived at a decision. Moreover, the debates tend to attract viewers who have an abiding interest in politics and are mostly party loyalists. Instead of the debates affecting who they will vote for, their party loyalty affects who they believe won the debates. For example, in a CNN poll after one of the 2008 debates, 85 percent of Democrats thought that Obama had won, but only 16 percent of Republicans agreed.
The impact of debates is also limited because the candidates are fairly evenly matched. Each candidate will have read a thick stack of briefing papers and rehearsed extensively. They will stick to their message and won’t be easily rattled. One candidate’s argument will be immediately countered by the other’s. Perhaps one candidate may appear more comfortable than the other. Perhaps one may momentarily slip up while the other does not. But the differences in their respective performances will be small. Candidates sometimes try to lower expectations of their own debate performance by claiming that they are just humble, plainspoken folks while their opponents are the second coming of Cicero. But Erikson and Wlezien’s analysis shows that across the series of debates in any given election year, the candidates tend to fight to a draw—much as one would expect two equally matched candidates to do.