Where my statisticians at?

Ha! You ended a sentence with a preposition and left out the verb. Won't your teachers be proud!

Ha! You ended a sentence with a preposition and left out the verb. Won’t your teachers be proud!

Pardon me while I stop shaking from having posed such a grammatically incorrect question. I am an editor, after all.

Whew! OK, so I’m calling on statisticians and research geeks alike to answer this question:

“Are numbers biased?”


 

I ask because a reader decided that my column today was biased to the extreme left, partly because of this passage:

I think it’s more important to check out their votes on specific issues, but the percentage of votes with the party is a good measure of moderation. In that vein, let’s check out OpenCongress’ ratings of party-line voting for Arkansas’ delegation.

In the House, Steve Womack and Tim Griffin lead the pack in partisanship, voting with the party 98 percent of the time. Rick Crawford is right behind at 97 percent. Tom Cotton brings up the rear with 95 percent. In the Senate, the most partisan of our two senators is John Boozman, who voted with his party 88 percent of the time. Mark Pryor, currently up for re-election, was at 82 percent.

Cotton, who is running for Pryor’s Senate seat, ranks 99th among 233 Republicans in the House for votes with the party, so while his votes make him the most moderate of our House delegation, in the context of the whole House, he’s not so moderate.

Of 51 Democrats in the Senate, Pryor ranks 50th in votes with his party, more often crossing the aisle and voting with Susan Collins of Maine than with Majority Leader Harry Reid. I’d call that pretty moderate, but that’s not what you see in the campaign ads.

My take is that I probably should have added the word “comparatively” in front of moderate, but other than that, I simply reported the numbers and rankings. I did comment on how moderate Cotton and Pryor are in the context of party-line voting, but I’m at a loss to see extreme-left views in that. Then again, perhaps if I were extreme right and/or more prone to bring my own bias …

I’ve always been of the opinion that statistics themselves, i.e., raw numbers and other data, are not inherently biased. Where bias usually enters the picture is in how those data are interpreted. Am I wrong? (Really, I want to know.)

So, dear readers, let me know your thoughts in the comments. I may include them in my column next week. I’ll try to keep snark to a minimum … at least about your answers. 😉

Advertisements