Tuesday, April 17, 2018

What??? No Pulitzer for PolitiFact in 2018?

We're not surprised PolitiFact failed to win a Pulitzer Prize in 2018.

The Pulitzer it won back in 2009 was a bit of a fluke to begin with, stemming from prize submissions in the public service category and arbitrarily used by the Pulitzer board to justify the prize for "National Reporting."

Since the win in 2009, PolitiFact has won exactly the same number of Pulitzer Prizes that PolitiFact Bias has won: Zero.

So, no, we're not surprised. But we think Hitler might be, judging from his reaction when PolitiFact failed to win a Pulitzer back in 2014.

Friday, April 13, 2018

PolitiFact continues to botch the gender pay gap

We can depend on PolitiFact to perform lousy fact-checking on the gender wage gap.

PolitiFact veteran Louis Jacobson proved PolitiFact consistent ineptitude with an April 13, 2018 fact check of Sen. Tina Smith (D-Min.), Sen. Al Franken's replacement.

Sen. Smith claimed that women earn only 80 cents on the dollar for doing the same jobs as men. That's false, and PolitiFact rated it "Mostly False."

That 80-cents-on-the-dollar wage gap is calculated based on full-time work irrespective of the job type and irrespective of hours worked once above the full-time threshold. The figure represents the median, not the average.

But isn't "Mostly False" a Fair Rating for Smith?

Good question! PolitiFact noted that the figure Smith was using did not take the type of job specifically into account. And PolitiFact pointed out that Smith made a common mistake. People often fail to mention that the raw wage gap figure doesn't take the type of job into account.

PolitiFact's Jacobson doesn't precisely spell out why PolitiFact finds a germ of truth in Smith's statement. Presumably PolitiFact's reasoning matches that of its earlier ratings where it noted that the wage gap statistic is accurate except for the part about it applying to equal work. So it's true except for the part that makes it false, therefore "Mostly False" instead of "False."

Looking at it objectively, however, it's just plain false that women earn 80 cents on the dollar for doing the same work. Researchers talk about an "unexplained gap" after taking various factors into account to explain the gap, and the ceiling for gender discrimination looks like it falls to around 5 percent to 7 percent.

Charitably using the 7 percent figure as the ceiling for gender-based wage discrimination, Smith exaggerated the gap by 186 percent. It's likely the exaggeration was far greater than that.

For comparison, when Bernie Sanders said 40 percent of U.S. gun sales occur without background checks, PolitiFact gave him a "False" rating for exaggerating the right figure by 90 percent.

The Ongoing Democratic Deception PolitiFact Overlooks

If a Democrat describes the 80 percent raw pay gap accurately, why not give it a "True" rating? Or at least "Mostly True"?

Democrats tend to trot out the raw gender pay gap statistic while proposing legislation that supposedly addresses gender discrimination. By repeatedly associating the raw wage gap with the issue of wage discrimination, Democrats send the implicit message that the raw wage gap describes gender discrimination. It uses the anchoring bias to mislead the audience about the size of the pay gap stemming from gender discrimination.

Democrats habitually use "Equal Pay Day," based on the raw wage gap, to argue for equal pay for equal work. But the raw wage gap doesn't take the type of job into account.

Trust PolitiFact not to notice the deception.

Fact checkers ought to assist in making Democrats clarify their position. Are Democrats in favor of equal pay regardless of the job or hours worked? Or do Democrats believe the demands for equal pay apply only to matters of gender discrimination?

If the latter, Democrats' continued use of the raw wage gap to peg the date of its "Equal Pay Day" counts as a blatant deception.

If the former, voters deserve to know what Democrats stand for.


It amused us that Jacobson directly referenced an earlier PolitiFact Florida botched treatment of the gender pay gap.

PolitiFact Florida couldn't figure out that claiming the gap occurs "simply because she isn't a man" is equivalent to claiming the raw gap is for men and women doing the same work. Think about it. If the gap occurs "simply because she isn't a man" then the reason for the disparity cannot be because she is doing different work. Doing different work would be a factor in addition to her not being a man.

PolitiFact Florida hilariously rated that claim "Mostly True." We wrote about it on March 14, 2017.

Fact checkers. D'oh.

Thursday, April 12, 2018

Not a Lot of Reader Confusion X: "I admit that there are flaws in this ..."

So hold my beer, Nelly.

We say that PolitiFact's graphs and charts, including its PunditFact collections of ratings for news networks, routinely mislead readers. But PolitiFact Editor Angie Drobnic Holan says there isn't much reader confusion.

Quite a few folks apparently have no clue at all that PolitiFact's charts and graphs lack anything like a scientific basis. Others know that something isn't right about the charts and graphs but against all reason find some value in them anyway.

PolitiFact itself would fall in the latter camp, based on the way it uses its charts and graphs.

So would Luciano Gonzalez, writing at Patheos. Gonzalez listened to Speaker of the House Paul Ryan's speech announcing his impending retirement and started wondering about Ryan's record of honesty.

PolitiFact's charts and graphs don't tell people about the honesty of politicians because of many flaky layers of selection bias, but people can't seem to help themselves (bold emphasis added):
I decided after hearing his speech at his press conference to independently check if this House Speaker has made more honest claims than his predecessors. To did this I went to Politifact and read the records of Nancy Pelosi (House Speaker from January 4th 2007-January 3rd 2011), John Boehner (House Speaker from January 5th 2011-October 29th, 2015), and of course of the current House Speaker Paul Ryan (October 29th 2015 until January 2019). I admit that there are flaws in this, such as the fact that not every political claim a politician makes is examined (or even capable of being examined) by Politifact and of course the inherent problems in giving political claims “true”, “mostly”, “half-true”, “mostly false”, “false”, & “pants on fire” ratings but it’s better than not examining political claims and a candidate’s level of honesty or awareness of reality at all.
If we can't have science, Gonzalez appears to say, pseudoscience is better than nothing at all.

Gonzalez proceeds to crunch the meaningless numbers, which "support" the premise of his column that Ryan isn't really so honest.

That accounts for the great bulk of Gonzalez's column.

Let's be clear: PolitiFact encourages this type of irresponsible behavior by publishing its nonsense graphs without the disclaimers that spell out for people that the graphs cannot be reasonably used to gauge people's honesty.

PolitiFact encourages exactly the type of behavior that fact checkers ought to discourage.

Monday, April 2, 2018

PolitiFact Bias Fights Fact Checker Falsehoods

A December 2017 project report by former PolitiFact intern Allison Colburn of the University of Missouri-Columbia School of Journalism made a number of misleading statements about PolitiFact Bias. This is our second post addressing that report. Find the first one here.

A blog, PolitiFactBias.com, devotes itself to finding specific instances of PolitiFact being unfair to conservatives. The blog does not provide analysis or opinion about fact-checks that give Republicans positive ratings. Rather, it mostly focuses on instances of PolitiFact being too hard on conservatives.
We find all three sentences untrue.

Does PolitiFact Bias devote itself to finding specific instances of PolitiFact being unfair to conservatives?

The PolitiFact Bias banner declares the site's purpose as "Exposing bias, mistakes and flimflammery at the PolitiFact fact check website." Moreover, the claim is specious on its face. After the page break we posted the title of each PolitiFact Bias blog entry from 2017, the year when Colburn published her report. The titles alone provide strong evidence contradicting Colburn's claim.

PolitiFact Bias exists to show the strongest evidence of the left-leaning bias that a plurality of Americans detect in the mainstream media, specific to PolitiFact. As such, we look for any manifestations of bias, including patterns in the use of words, patterns in the application of subjective ratings, biased framing and inconsistent application of principles.

Does PolitiFact Bias not provide analysis or opinion about about fact-checks that give Republicans positive ratings?

PolitiFact Bias focuses its posts on issues that accord with its purpose of exposing PolitiFact's bias, mistakes and flimflammery. Our focus by its nature is technically orthogonal to PolitiFact giving Republicans positive ratings. And, in fact, PolitiFact Bias does analyze cases where Republicans received high ratings. PolitiFact Bias even highlights some criticisms of PolitiFact from the left.

We simply do not find many strong criticisms of PolitiFact from the left. There are plenty of criticisms of PolitiFact from the right that we likewise find weak.

Does PolitiFact Bias "mostly focus" on PolitiFact's harsh treatment of conservatives?

PolitiFact Bias recognizes the subjectivity of PolitiFact's "Truth-O-Meter" ratings. PolitiFact's rating system offers no dependable means of objectively grading the truth value of political statements. For that reason, this site tends to avoid specifically faulting PolitiFact's assigned ratings. Instead, PolitiFact Bias places its emphasis on cases showing PolitiFact inconsistency in applying its ratings. In two similar cases where a Democrat received a positive rating and a Republican received a lower rating it might be the case that PolitiFact went easy on the Democrat.

That said, the list of post titles again shows that PolitiFact Bias produces a great deal of content that is not focused on showing PolitiFact should give conservatives more positive ratings. Holan's statement jibes with Colburn's false statement about the focus at PolitiFact Bias.

Why the misleading claims about PolitiFact Bias?

As far as we can tell, the entire evidence Colburn used in her report's judgment of PolitiFact Bias came from her interview with PolitiFact Editor Angie Drobnic Holan:
I'm just kind of curious, there's the site, PolitiFactBias.com. What are what are your thoughts on that site?

That seems to be one guy who's been around for a long time, and his complaints just seem to be that we don't have good, that we don't give enough good ratings, positive ratings to conservatives. And then he just kind of looks for whatever evidence he can find to support that point.

Do you guys ever read his stuff? Does it ever worry you?

He's been making the same complaint for so long that it has tended to become background noise, to be honest. I find him just very singularly focused in his complaints, and he very seldom brings up anything that I learn from. But he's very, you know, I give him credit for sticking in there. I mean he used to give us, like when he first started he would give us grades for our reporting and our editing. So it would be like grades for this report: Reporter Angie Holan, editor Bill Adair. And like we could never do better than like a D-minus. So it's just like whatever. What I find is it's hard for me to take critics seriously when they never say we do anything right. Sometimes we can do things right, and you'll never see it on that site.
Note that in Holan's response to Colburn's first question about PolitiFact Bias she suggests the site focuses on PolitiFact not giving enough positive ratings to conservatives.

Are Colburn and Holan lying?

PolitiFact Bias co-editor Jeff D. has used the PolitiFact Bias Twitter account to charge Colburn and Holan with lying.

The charge isn't unreasonable.

Colburn very likely read the PolitiFact Bias site to some extent before asking Holan about it. Even a cursory read ought to have informed a reasonable person that Holan's description of the site was slanted at best. Yet Holan's description apparently underpinned Colburn's description of PolitiFact Bias.

Likewise, Holan's familiarity with the PolitiFact Bias site ought to have informed her that her description of it was wrong and misleading.

When a person knowingly makes a false or misleading statement, it counts as a lie. Colburn and Holan were both very likely to have reason to know their statements were false or misleading.

We're pondering a second post pressing the issue still further in Holan's case.