PolitiFlub: Tom Coburn and the PolitiFact corrections game

When we make a mistake, we correct it and note it on the original item. If the mistake is so significant that it requires us to change the ruling, we will do so.
--Principles of PolitiFact and the Truth-O-Meter

Pity poor PolitiFact.  They knoweth not what they do.

PolitiFact tried to make good on its corrections policy this week by updating an item on Tom Coburn (R-Okla.).  Coburn, in the context of drawing attention to the increase of government spending, claimed the government has doubled in size since 2001.

PolitiFact conclusion version one:

Coburn is entirely right about federal outlays doubling over 10 years. And if you use one of two plausible measures, he’s right on the growth since Obama took office. But using another plausible measure, the growth of government under Obama was smaller. On balance, we rate Coburn’s statement Mostly True.

PolitiFact addicts may note that PolitiFact has broken its rule of dividing complex statements into individual claims for purposes of Truth-O-Meter ratings:

PolitiFact writers and editors spend considerable time deliberating on our rulings. We always try to get the original statement in its full context rather than an edited form that appeared in news stories. We then divide the statement into individual claims that we check separately. For example, a Bill Richardson TV ad produced two claims. (We only make Truth-O-Meter rulings on those individual claims. We don't make them in our articles because they often summarize multiple Truth-O-Meter items that had different rulings.)

PolitiFact's principles often seem more like ... guidelines.  They "only" make Truth-O-Meter rulings on the individual claims except when they don't.

But back to Coburn and PolitiFact conclusion version two:

We don’t think that Coburn was entirely wrong to use non-inflation-adjusted dollars, but it would have been better if he -- and we -- had adjusted for inflation. Doing it this way still means there was a significant increase in government over the period he studied, but not as high as he had indicated. On balance, we have decided to lower our ruling from Mostly True to Half True.

It would be pretty easy to accept PolitiFact's revised rating of Coburn if was consistent with PolitiFact's usual practice.  Sure, Coburn was probably using his doubling reference kind of like a "large as a football field" type reference to simply illustrate the growth in terms the average listener could easily grasp.  His point, as PolitiFact appears to note, was simply that the government has grown considerably in the past decade in terms of spending.


Afters

I got to wondering about other instances of doubling in PolitiFact lore, such as this one from PolitiFact Texas featuring the Democratic Party's gubernatorial challenger, Bill White.

"Debt has almost doubled in Austin under Gov. Perry," White said. "They think you will not notice this!"

To my admitted surprise, PolitiFact Texas did the calculations while adjusting for inflation.   There was no calculation for the grown of Texas' debt as a percentage of GDP, however.

Graphic credit:  mapzones.org

Using the above chart, Texas' debt as a percentage of GDP went from 1.9 percent for the year 2000 to about 3 percent (year 2009 GDP obtained from a table here).  That's about a 58 percent increase--not double, in other words.  The "Truth-O-Meter" rating?  "True."  And that's without any apparent consideration of White's dubious underlying point (Governor Perry's fault?).

On the other hand, the national PolitiFact operation neglected to calculate Democrat Chris Van Hollen's doubling claim regarding insurance premiums using inflation-adjusted dollars. Observe how the adjustment might have affected PolitiFact's analysis of Van Hollen.  Before adjusting for inflation:

In 2000, average annual premiums for single people were $2,471, a number that rose to $4,242 by 2006 — a 72 percent increase.

After adjusting for inflation:

In 2000, average annual premiums for single people were $2,892 in inflation-adjusted dollars, a number that rose to $4,242 by 2006 — a 47 percent increase.

Van Hollen's "Truth-O-Meter" rating as it stands today, with no correction for inflation?  "Mostly True."  Though I've used the percentage figure least favorable to Van Hollen as the example, the rest of his percentages would also have hovered near 50 percent for his doubling claim after an adjustment for inflation.  Compare Van Hollen's 47 percent figure to Coburn's 61 after inflation adjustments.

More?

President Obama received a "False" for a claim about doubling exports without the added damage from an inflation adjustment.  Missed your shot at "Pants on Fire," buddy, if it was a question of a few percentage points!

PolitiFact Oregon used inflation-adjusted dollars in giving Republican Scott Bruun a "Mostly False" rating for a claim eerily similar to Coburn's--and in the process inspired another "PolitiFlub" post.

This check on a Hillary Clinton claim has barely any research at all attached, and certainly no mention of an adjustment for inflation.  Why not just give her a "True" rating?

The bottom line?  The fact checks from PolitiFact often fail to bear up under close scrutiny.