AK-Sen: In Defense of Math

Awhile back, the old mainframe down at SSP Labs told us that not only that absentees and provisionals wouldn’t be enough to help Lisa Murkowski overcome her then 1,668-vote deficit, but that they’d be likely to actually hurt her and drop her further behind, to the tune of 341 votes.

Three weeks after election day, we finally have some certainty, as the results are now certified. Drumroll, please:






















Election Day Final
Joe Miller 47,027 55,878
Lisa Murkowski 45,359 53,872
Margin 1,668 2,006

With all the final results, Miller’s margin on Murkowski increased from 1,668 to 2,006. An increase of … 338!

Now, of course, things aren’t that simple, and frankly, we’re not that good. In my input, we’d assumed 7,500 absentees and 5,000 provisionals, when in actuality, there were 17,364 additional ballots. I don’t have those broken down between provisionals and absentees, but applying that figure post hoc, that would have yielded an estimate of an increase in Miller’s margin of between 410 and 516.

But indeed, absentees and provisionals were not friendlier to Murkowski – she lost by 1.81% among election day ballots and early votes; she lost by 1.95% among provisionals and absentees. (We’d estimated she’d lose by 2.36% among provisionals and 2.97% among absentees).

No models are perfect, and they remain exactly that – models, dependent on anorexia! er…proper inputs and proper assumptions; we’ll continue to put them forth to the best of our ability. Moving forward, even as the news may get even worse (or even better, if you’re one of our many newfound GOP-leaning readers), we hope to bring the same solid news and analysis (that we assume) you come here to enjoy!

P.S. I’m no salesman, but since I’ve already gotten wet in the self-congratulatory rain, I might as well go swimming: we’re a mere 5 followers short of 3,000 on Twitter. Wink wink. Nudge nudge.

P.P.S. We have a Facebook page too.

7 thoughts on “AK-Sen: In Defense of Math”

Comments are closed.