honoluluadvertiser.com

Sponsored by:

Comment, blog & share photos

Log in | Become a member
The Honolulu Advertiser
Posted on: Wednesday, December 31, 2008

Heads or tails, it's a toss-up in Minnesota

By Jerry Burris
Advertiser Columnist

The ongoing confusion over who won the election for U.S. Senate from Minnesota shines a bright spotlight on one of the least understood secrets of modern electoral politics.

That secret is that results of elections, although reported in discrete and apparently final numbers, are really approximations at best. The count is never as conclusive as voters and election officials would wish us to believe.

This is not because of inept election officials or faulty machines, as many would like to think. It is the result of the inevitable imperfection of counting numbers when they rise into the thousands or millions, as is the case in Minnesota.

The process of counting votes — and Hawai'i has a good record on this front — is amazingly accurate in the aggregate. If the count declares Candidate A the winner, then most likely he or she would be the winner no matter how many times the votes are counted. It's a good approximation of reality.

But here's the point: Each of those counts — no matter how many times they are conducted — is likely to come up with slightly different results.

Not enough to matter statistically, perhaps, but meaningful when the winner is decided by one more vote than the loser.

Simply put, when numbers are of such magnitude, it is all but impossible to come up with precisely accurate counts.

In the case of Minnesota, the difference between Democrat Al Franken and Republican Norm Coleman at the moment is something like 47 votes out of more than 2.9 million ballots counted. Mathematicians will tell you that is an insignificant difference; by any reasonable measure the two are tied.

But this is politics, not mathematics. We need winners and losers.

So the counting will go on and on.

It is a sure bet that if they start the counting over again, they would come up with slightly different results — not different enough to matter to a statistician but different enough to matter to the two candidates involved.

In Minnesota, they should simply declare the race too close to call and let the candidates flip a coin to decide who will become the next senator.

They will never settle this through the counting process. In fact, there is a law to this effect and the two candidates should use it.

Not to put too fine a point on it, but this illustrates once again the importance of getting out the vote and making sure that everyone who is entitled to vote actually does so. The greater the number of votes, the less opportunity there will be for matters to be settled by statistical chance.

Jerry Burris' column appears Wednesdays in this space. See his blog at blogs.honoluluadvertiser.com/akamaipolitics. Reach him at jrryburris@yahoo.com.