[link|http://www.nytimes.com/2006/11/06/opinion/06conley.html?_r=1&oref=slogin|NY Times OpEd by Dalton Conley]:

[...]

In an era of small town halls and direct democracy it might have made sense to rely on a literalist interpretation of \ufffdmajority rule.\ufffd After all, every vote could really be accounted for. But in situations where millions of votes are cast, and especially where some may be suspect, what we need is a more robust sense of winning. So from the world of statistics, I am here to offer one: To win, candidates must exceed their rivals with more than 99 percent statistical certainty \ufffd a typical standard in scientific research. What does this mean in actuality? In terms of a two-candidate race in which each has attained around 50 percent of the vote, a 1 percent margin of error would be represented by 1.29 divided by the square root of the number of votes cast.

Let\ufffds take the Washington gubernatorial race in 2004 as an example. After a manual recount, Christine Gregoire was said to have 1,373,361 votes, 48.8730 percent, while her Republican rival, Dino Rossi, garnered 1,373,232, or 48.8685 percent (a third-party candidate got 63,465 votes). That\ufffds a difference of only 129 votes, or .0045 percent. The standard error for a 99 percent certainty level was 0.078 percentage points. Since Ms. Gregoire\ufffds margin of victory didn\ufffdt exceed this figure, under this system she wouldn\ufffdt be certified as the victor.

If we apply the same methodology to Bush v. Gore in 2000, the results are equally ambiguous. The final (if still controversial) vote difference for Florida was 537 (or .009 percent). Given Florida\ufffds vote count of 5,825,043, (excluding third party votes) this margin fails to exceed the 99 percent confidence threshold. New Mexico, which Al Gore won by 366 votes out of a much smaller total, is also up for grabs in this situation.

So what should we do in such cases, where no winner can be declared with more than 99 percent statistical certainty? Do the whole shebang all over again. This has the advantage of testing voters\ufffd commitment to candidates. Maybe you didn\ufffdt think the election was going to be as close as it was, so you didn\ufffdt vote. Well, now you get a second chance.

[...]

I like it. It seems to be a simpler change than something like [link|http://en.wikipedia.org/wiki/Instant-runoff_voting|instant runoff voting] or [link|http://en.wikipedia.org/wiki/Cumulative_voting|cumulative voting] (both of which also offer some advantages) but it would help ensure that a small number of questioned ballots can't sway the will of the majority (or plurality). Of course, finding a way to do the second election that encourages high turn-out may be difficult but, as he points out, many states already require run-offs in some elections.

Cheers,
Scott.

It's the tyranny of large numbers. It's unreasonable to expect that, even with such audits and paper receipts, that reproducible perfect counts that reflect the will of the people will always happen.

E.g. What about absentee ballots?

Requiring that the results have to differ by a reasonable margin when the sides are very close seems to me to be smart.

Cheers,
Scott.