Still, many political observers seem to struggle to understand how hindsight bias is clouding their judgments and the media’s coverage. It can be difficult to imagine the stories we would be seeing now if Romney had won, for instance. That’s why it’s instructive to consider Politico’s premortem on how an Obama loss would be interpreted, which seems like a plausible counterfactual account of the stories that would now be circulating. (They also wrote a Romney premortem that didn’t anticipate the emphasis that would be given to Obama’s ads and the ground game.) Or consider the story of the 2004 election. According to an American Journalism Review article, reporters relying on flawed exit polls ended up writing stories “explaining” why Kerry won that had to be replaced by alternate accounts recounting how Bush won a narrow victory:

Like their TV counterparts, many newspaper reporters saw early exit polls, and some crafted preliminary stories relying on those numbers.

Jack Torry, a Washington reporter for Ohio’s Columbus Dispatch, helped write two different analysis stories on election night. The first, never published, was based on late afternoon exit polls and explored how Kerry won the election, a “referendum on President Bush.”

“The more I kept checking with Republican sources who I really do trust, the more I began to wonder: Could those exit polls be right?” Torry says. He and his colleagues scrapped the first piece around 9:30 p.m. Their second analysis, which actually appeared in the paper, examined why the race was so tight.

Before that election, Newsweek went so far as to promote two different versions of a post-campaign book about how either Bush or Kerry won before the election had even happened—check out this amazing image:

A similar process of retrospective justification would have likely ensued among observers had Kerry’s reported lead held up.

Of course, it may actually be true that Obama won this election because of his superior ground game or devastating barrage of negative advertising, but convincing evidence does not yet exist to support either claim. Obama performed about as well as we would have expected given the political and economic fundamentals of the race. Until proven otherwise, little credence should be given to post-election narratives that are constructed after the fact to “explain” the outcome we have observed.


More in United States Project

Four stories to follow in Virginia

Read More »

Brendan Nyhan is an assistant professor of government at Dartmouth College. He blogs at and tweets @BrendanNyhan.