Archive for September 2012
Quick post, but something that has been bothering me:
Fact checkers like PolitiFact have been coming under fire from all corners recently, some of it very well deserved. But many conservatives have jumped on the fact that more statements made by Republicans or conservatives are rated “false” or “pants-on-fire” as opposed to statements by Democrats or liberals. Look!, they say, clear evidence that the fact checkers are biased!
Here are a few examples:
“In July you printed a chart with two years of PolitiFact Ohio results. It showed Democrats with 42 ratings of Mostly False, False or Pants on Fire, while the Republicans had a total of 88 in those categories. Doesn’t that prove you guys are biased?”
And, slightly more sophisticatedly:
“A data-driven analysis of PolitiFact Florida’s 554 rulings on statements made by individuals appears to show a clear bias against Republicans and in favor of Democrats. As the truthfulness of a statement increases, so does the percentage of Democratic claims included in PolitiFact Florida’s rating.
… This dynamic appears to be a textbook example of what statisticians call ‘selection bias.’”
(This is followed by two cherry picked examples of how and whether PolitiFact chose to review certain statements.)
Here’s the thing these two posts neglect to mention: not reporting an exact 50/50 split is only evidence of bias if the split is in reality 50/50. Read the rest of this entry »