Fox News Posted a Poll That Doesn't Add Up — Again

Fox News Posted a Poll That Doesn't Add Up — Again

The news: Boy, this is odd. Fox News posted the following poll during July 30's edition of Special Report, which seems to indicate that 110% of voters in Senate battleground states disapprove of the president's job performance.

Quality work, Fox.


(H/T to Media Matters for pointing it out.)

If you believe this poll, I've got a bridge to sell you in Wasilla. It goes nowhere, which is probably farther than where this methodology will get you in 12 heated Senate races. The full clip is below and it's plainly obvious that Fox didn't even make an effort to represent the poll accurately.


Here's how Democracy Corps, the firm that actually conducted the polling, presented the poll. It is still awful news for Democrats, but it doesn't try to count opinions twice.


The background: This would be embarrassing if it happened to any channel but Fox News, which clearly doesn't care.

The network puts a premium on promoting a right-wing agenda and isn't above pushing a comforting image of constant disaster among the Obama administration to placate their ultraconservative base.

It's not like this is even close to the first time. Mediaite's Matt Wilstein caught this misleading chart in March on America's Newsroom, which portrays Obamacare signups as disastrously under its target when in fact the administration solidly hit its revised goals — as shown on another Fox program the previous day.

Here's the misleading one, compared to its accurate cousin.



They actually lied about Obamacare another time, notes Dan Brooks:


Here's a 2012 approval poll, cited by POLITICO's Byron Tau, that also adds up to more than 100% — and since this one isn't double-counting, it's just plain wrong rather than misleading.


From the expiration of Bush's tax cuts ...

... to the deficit ...


... to whether global warming is a myth ...


... to unemployment rates ...



... the charts at Fox News are clearly meant to paint a misleading picture for anyone who isn't paying close attention or sucks at math. Conveniently, it's always biased towards conservatives.

They're probably not mistakes: Simply Statistics' Jeff Leek analyzed a series of misleading FNC charts in 2012 and found that the charts didn't appear to be simple graphical errors or incompetence. Instead, they used a series of techniques like truncating the y-axis, presenting multiple-option polls as if they were single-option, screwing around with the units of comparison, messing with the magnitude of units at different x-values or spacing various points on the x-axis out just enough to distort infographics.

Leek, who develops quantitative tools for genomic data, concluded the techniques involved are clever enough that they're not accidents. In other words, some right-wing statistician at Fox News is laughing all the way to the bank.