Corruption Smackdown: The Graph
Andy suggests a graph to replace Lee’s table below. Here is one:
This takes the rankings out of the graph. I’ve noted before that rankings often distort data. This plot shows more clearly that Illinois is really just one of a pack behind the frontrunners Louisiana and Mississippi, and that a lot of states are essentially tied.
Comments
To my mind, convictions per resident is the wrong measure to use. If we want to know how corrupt a state government is, we shouldn’t be comparing the number of convictions to the number of people who live in the state. Rather, we should compare the number of convictions to the size of the public sector. So we should really be observing convictions per number of elected officials, or as a proportion of total number of people employed by the state.
Posted by: Jeff Lazarus | December 16, 2008 11:14 PM
Where’s Alaska in the list? I think they’ve proven no immunity to corruption…
Posted by: Tom Terry | December 17, 2008 01:12 AM
Jeff: That point was already made in response to my original post. Please collect those data, do the analyses, and report back to us. :-)
Tom: As noted in the comments to my original post, just the 35 most populous states were included in the analysis.
John Sides: So I do all the hard work and you barge in at the last minute to bask in the glory. It figures. But I’ve learned something important from your post: Next time, instead of posting real, live, good-solid-American numbers, I’ll ask you to convert the numbers into cute little dots and post them on my behalf.
Posted by: Lee Sigelman | December 17, 2008 09:20 AM
A map would be good, to show the regional patterns.
Also, it’s clear that Oregon is the most corrupt . . . they’re so damn good at it that they don’t even get caught!
Posted by: Andrew Gelman | December 17, 2008 11:08 AM