Tom Wood of Ohio State University and Ethan Porter of George Washington University published their paper in April to London-based journal Electoral Studies, claiming the political satirist's absence on the satirical news show led to a 1.1 percent increase in Trump’s vote share.
The researchers collected rating for the show across the U.S. and county-level voter turnout for 2012 and 2016. They theorized that the change in hosts from Stewart, who mocked Trump nightly, to Trevor Noah, led to a dip in ratings. The result was Trump earning around 1 percent more of the votes in each county.
"Our results make clear that late-night political comedy can have meaningful effects on presidential elections," the paper stated.
Readers quickly alerted the pair to an error in their findings. After checking their work, the pair realized a computational error altered their conclusion.
“Upon closer investigation ... we discovered we had made a computational error in aggregating years of ratings data,” Wood tweeted on May 10. “After fixing this error, the corrected regression results do not implicate changes in the Jon Stewart and Stephen Colbert audiences as important correlates of Donald Trump’s 2016 presidential vote.”
The pair notified the journal and asked that the paper be withdrawn, to which it agreed. They even received praise on Twitter for publicly acknowledging the mistake.