I have tended to be a big fan of Open Secrets, a website by a lefty group that shows how money is spent in politics. The nice thing is that they provide data. And the nice thing about data is that it gives you good apples-to-apples comparisons. Based on data, you can argue things like "the largest donors and lobbyists in Wisconsin are the teachers unions" and have something to back that up with.
So I was really disappointed to read this recent piece by Michael Beckel at Open Secrets entitled "Union Muscle Eclipsed by High-Profile Conservative Groups During 2010 Election". They reviewed the publicly disclosed spending information and concluded that the unions spent less money than American Crossroads, the Chamber, et al. in the 2010 cycle. In particular, they found that the unions spent $46.7m while business groups spent $97m or so.
There's a catch though. AFSCME, the largest of the public employee unions, told the New York Times that they spent $91m. That number was actually up $3.5m from four days before then when a union representative told the Wall Street Journal, "we're the big dog, but we don't like to brag." Open Secrets claimed that AFSCME only spent $12.6m, less than 1/7th of the total amount that the union says that they spent.
Now I am not saying that the apples-to-apples study of disclosed data to disclosed data isn't valuablein some cases. But when your study understates the expenditures of one organization by a factor of 7 you have to think that you are barking up the wrong tree in how you are collecting your data. It isn't so much that your methodology is bad. It is that it is irrelevant and misleading.
Of course, the guy who did the analysis used to work for the extremely liberal Mother Jones. Maybe he just knew what answer he wanted and picked the parameters to hit it.