Get out your wizard hats! It’s starting to sound like campaign season again.

Just as political reporters wanted to tell evidence-free stories about Mitt Romney’s “momentum” back in October, many media members are now being swept up in the Obama administration scandal narrative and making claims about the president’s declining standing with the public that aren’t yet supported by the data.

On Thursday, the Quinnipiac University Polling Institute issued a press release suggesting that Obama’s approval rating had been damaged by the controversy over the Internal Revenue Service’s treatment of conservative groups. According to their May 22-28 poll, Quinnipiac wrote, “Obama gets a negative 45 - 49 percent job approval rating, compared to 48 - 45 percent positive in a May 1 survey… conducted before the IRS allegations surfaced.”

Given that the poll results each had a 2.6% margin of error, reporters should have been cautious about making too much of this shift, especially given the evidence that only the most significant scandals have much effect on presidential approval. But the suggestion that Obama’s approval rating had been damaged by the controversies was quickly seized on by news-hungry journalists and commentators. In an online interview Friday, Politico’s Patrick Gavin and Fox News Sunday host Chris Wallace took the bait:

GAVIN: Chris, we saw a new Quinnipiac poll out this week that shows a slight downtick in the president’s approval ratings. Is it safe to say that the IRS, Benghazi, DOJ scandals might be taking a hit on the president’s approval rating?

WALLACE: Well, it would seem that’s the one thing - I mean, the economy’s doing pretty well, so you gotta figure the one development over the last month has been the fact that the president is getting hit on all of these things, and I think particularly the IRS…

Wallace, in turn, asked virtually the same question to Republican National Committee chairman Reince Priebus on Fox News Sunday over the weekend:

WALLACE: All right. Let’s turn to the scandals that President Obama is dealing with right now. According to a new Quinnipiac poll, and let’s put it up on the screen, 45 percent now approve of the job the president is doing while 49 percent disapprove. That is a swing of minus 7 points from just a month ago.

With the economy, and a lot of the numbers in the economy generally doing better, do you think that drop in the president’s approval rating, minus 7 points in a month, is because of the scandals?

(Priebus rejected the premise that the economy is doing better, but agreed that the storyline was “taking its toll.”)

Wallace and his interlocutors weren’t the only media figures to make that connection. Fox News White House correspondent Ed Henry (“A new Quinnipiac poll shows those controversies are taking some toll on the president”), MSNBC host Chris Matthews (“A trio of administration controversies, if you can call them that, and a weak economic message, no surprise that Obama is finally slipping in the polls now”), and Bloomberg’s Jonathan D. Salant (“The spotlight on the controversies has taken a toll on Obama’s standing”) all offered similar takes after the poll’s release.

But there’s a problem with those conclusions, as the political scientists Charles Franklin and Jonathan Bernstein pointed out: a broader look at the evidence shows little evidence of a decline in Obama’s approval rating in recent weeks. For instance, the current estimate from Huffington Post Pollster (where I sometimes blog) of Obama’s approval rating is 47.3%—virtually identical to the 47.8% estimate on May 9, the day before the IRS scandal began:

Indeed, the media narrative about the scandals and Obama’s falling standing with the public is based on a change in approval ratings over the past month that appears to be an outlier. When we zoom in closer to the data, we can see that the Quinnipiac polls (the dashed lines) show a much sharper increase in Obama disapproval and decrease in Obama approval than the overall polling averages (the solid lines):

It’s possible, of course, that Obama’s approval ratings will decline, but the evidence does not currently support such a claim. The fluctuations seen in Obama’s Quinnipiac approval ratings are likely attributable to random sampling error—the precise reason we use polling averages in the first place! As both CJR’s Greg Marx and I showed back in 2010 after the Deepwater Horizon oil spill in the Gulf of Mexico, the temptation to tell stories about how major events are affecting presidential approval using outlier polls is strong, but should be resisted.

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