There are two ways to approach being evenhanded: You can try to actually be evenhanded, which could mean that you find that one side is right and the other is wrong. Or you can strive for the appearance of being evenhanded, which means that you decide in advance that you're going to find that there's truth on both sides.
PolitiFact, a political factchecking project based in St. Petersburg, Florida, has been criticized for taking the latter approach. An item it posted yesterday (1/9/12) is further evidence of its preference for the appearance of evenhandedness over its reality.
The item addressed Rick Santorum's assertion in a January 4 town meeting that as a result of the 1996 welfare law, "Poverty levels went down to the lowest level ever for…one of the areas that had the highest level of poverty historically, which is African-American children." PolitiFact concluded that the statement was "Half True," since "Santorum is right that poverty rates declined after the reform's passage. But opinions differ on the primary cause."
As evidence that "opinions differ," the factcheckers turned to Robert Rector of the Heritage Foundation, best known for his argument that the poor aren't really poor because they have microwave ovens and the like. Unsurprisingly, since he works for a group set up explicitly to promote conservative ideas, he does indeed have the opinion that the 1996 welfare law caused a drop in child poverty. But does this opinion have any basis in fact?
PolitiFact allows him to make his case at length, but the gist of it is this: "Since welfare reform, the poverty rate among black children has fallen at an unprecedented rate from 41.5 percent in 1995 to 32.9 percent in 2004." And PolitiFact helpfully gives you a link to a U.S. Census chart that shows that those numbers are almost accurate. But looking at the numbers for yourself, you see that there's no indication that the 1996 law had anything to do with them: Poverty among black children peaked in 1992, at 46.3 percent, and declined steadily from then until 2001, when it hit a low of 30.0 before moving upward. 1996 does not seem to have impacted the poverty trajectory at all; a naive reading of the numbers would indicate that black child poverty goes up when someone named "Bush" is in the White House.
Here's a graph of child poverty by race from Mother Jones (9/29/11–by raw numbers, not percentages) that illustrates the utter unremarkability of 1996 for black child poverty:
PolitiFact goes on to give equal space, and equal rhetorical weight, to sources who say economic growth is actually what drove child poverty down in the '90s: "While Rector maintains that the economy played only a secondary role in reducing poverty, other groups says it's the main driver." But none of these sources directly rebut Rector's arguments, or point out how dubious it is to give a 1996 law credit for a decline that began four years earlier.
So it's true that "opinions differ" on whether the 1996 welfare lowered poverty for black children. A real factchecker would point out that the advocate for that opinion offers selective and misleading figures to back it up. But then, if you did point that out, you might look like you weren't being evenhanded.
(Thanks to Neil deMause for bringing PolitiFact's report to my attention.)