If you can think of any particular inaccuracies I'd love to see them. I'm sure there are some, but I haven't noticed any and I'm curious about them.
Some of the statements Politifact rates are indeed, as Kyl would put it, not intended to be factual statements. Again, though, I think that if you read their full analysis you'll generally end up better informed about the facts, even if the statements themselves pass beyond the factual. I certainly wouldn't consider what Politifact does to be more inflammatory than general reportage.
The Economist, in their analysis of Politifact's decision (
Politics and lies: Fact-checking the fact-checkers | The Economist) also points out that an inaccuracy is not necessarily a deliberate inaccuracy (although "lie" does imply intent to deceive).
Here is one example.
PolitiFact | Rand Paul says federal workers paid $120,000, private-sector workers only $60,000
Their argument that most people do not think of benefits on a par with salary is complete hokum, and rating Paul's statement a lie by claiming that most people would not think about benefits when hearing that is indicative of their bias in favor of the left in rating statements.
I think you and I agree on how Politifact works, we just disagree on the wording. What I said was:
I think that if you read their full analysis you'll generally end up better informed about the facts.
In other posts, I have said that Politifact's raw rankings (in the example you gave, "False") can be misleading if one does not read the analysis.
I think that both of these things are true in the given case. You seem to believe that federal employees earn roughly twice as much as private sector employees on average in total compensation. Someone who looked only at Politifact's rating might think they were saying that was false. Someone reading their analysis would see that they said
According to the Bureau of Economic Analysis, a federal statistics-gathering agency, federal worker compensation in 2009 averaged $123,049, which was double the private-sector average of $61,051. That's a gap of almost $62,000 -- and is pretty close to what Paul said on This Week.
So, as I said, I think someone who read their full analysis would end up better informed.
Incidentally, while I concede that Politifact's ratings can be misleading, I'm not sure there is a better way to do them. If one becomes excessively literal-minded, one has to rate such statements as "She [a Texas mother] told me [Bachmann] her daughter suffered mental retardation as a result of that vaccine" as true (rather than false, as Politifact in fact rated it) no matter how misleading they are.
(Regarding whether their argument is "hokum", I would say it is plausible but unsupported. I think most salaried workers, when asked how much they made in a year, would not include the monetary value of their benefits. Still, Politifact didn't prove that. My point, though, is that after reading their analysis you know they didn't prove it- they didn't manufacture proof.)