Leaving out specific items is an obvious way to work an agenda into it, but manipulating the groupings could do the same thing, and more subtly at that. With a little gerrymandering, the receipt could say pretty much anything.

>4 Budget calculations are based on budget allocations obtained from the Office of Management and Budget using historical tables from the FY2011 budget submission as well as the Congressional Budget Office website for such items as interest on the debt, Medicare, Medicaid, and Social Security.

My guess is that the real cause of "selected items" is that the number of things paid for by your taxes are obviously much larger than could fit on a short list, and has a long tail.

Well, of course that depends how you slice it. Break it down by department and it looks more like this:

http://en.wikipedia.org/wiki/File:Fy2010_spending_by_categor...

The twenty-something listed departments account for 97% of the budget (of course there's other government spending which isn't listed explicitly in the budget, I suppose).

This doesn't, of course, include stuff like "Smithsonian Museum", which are included on this list to make people feel warm and fuzzy about paying their taxes.

Somewhat disingenuous from a think tank called the "Third Way", which was originally intended to imply centrism. Still, it's a general law of NPR reporting that a left-wing think tank is called a "think tank" and a right-wing think tank is called a "right-wing think tank".

I don't really have a dog in this fight, but a quick search suggests you may be mistaken. [site:npr.org "right-wing think tank"] os 8 results; [site:npr.org "think tank"] has about 2380. Nothing funny in robots.txt.

But you never know, maybe there aren't that many transcripts on there or something...

Still, a bit more googling:

7 hits for "right-wing think tank" vs 4 hits for "left-wing think tank", so apparently they don't use either term very often.

1200 hits for "liberal think tank" vs 285 for "conservative think tank" vs 447 for "libertarian think tank" (all of which seem to be Cato) vs 6 for "moderate think tank".

3 for "Republican think tank", 345 for "Democratic think tank" and 57 for "non-partisan think tank"

I take your point, though. What you'd really need to do in order to analyze bias would be to check out "think tank" is used without an adjective and figure out what kind of think tanks those are.

The empirical evidence shows that, for example, NPR is more likely to interview a conservative economist about a story than a liberal. I think they go out of their way to defer to right-of-center politicians because they are so afraid of the perception that they're liberal.

In short, I think you're wrong.

To be meaningful, we'd need to compare the ratio of conservative economists on NPR to conservative economists in the population.