Over at Wired.com there is an interesting little post about how statistics are being used to calculate the likelihood that mistakes in counting at some localities are evidence that the overall election result should be invalidated. I’ve long thought that there must be a way of insuring the count is correct. Anyone who has tried to get an exact count of the exact number of items when there are a lot knows that the process is not as easy as it seems when you say, “just count them up.” I’m glad to see that that California is acknowledging that probability and statistics can help us get an idea of when there is reasonable cause to question the reliability of the results.
While I do think the move to computer polling should help with accuracy of recording and tallying the results (computers are much better than humans at simple rote tasks like this), I do worry about the inability to recount/recreate the votes should there be a software or hardware issue that causes the loss of information. Be that as it may, you can check out the article at the link above, but I’ll include the graphic explaining the math here. Congrats to those of you who, like me, are savvy enough to get a pretty good idea of what they are trying to do with this measure. The pieces are explained in the original piece. Just click on the image here, and scroll down.