Great Moments in Responsible Journalism...
The father of Brennan Hawkins is surrounded by his entire family on his porch with dozens of media people around him.
DAD: We would like to field questions from the media now. Our only request is that you do not ask Brennan any questions.
Off-screen Reporter: Brennan! How are you feeling?
DAD: I just asked you not to ask Brennan any questions.
Reporter: Oh; Dad -- How's Brennan doing?
Wednesday, June 08, 2005
I think everyone loves the drama of a close election. The image of diligent polling employees pulling vote slips out of a box and tallying the results one by one is made even more exciting when a winner is declared by just one vote. It's a beautiful image but not a very realistic one.
With the possible exception of a heated campaign for fourth grade class president, elections just aren't that tidy. The fact is that votes get lost, miscounted, etc. The reality is that in a largish election such as in the city of Seattle or the largest election, that of the President, there is no way to state with certainty what the exact number of votes for each candidate is. Ignoring for the moment the accusations of fraud just the logistics of aggregating all those votes is bound to cause some lack of exactness.
This 'lack of exactness' or 'statistical error' to use the parlance of our time gives rise to the possibility of a statistical tie even when one candidate has a nominally larger number of votes. It is a standard issue in statistical science that is resolved quite adequately by the science. I think it is time that we started including some of theory into practice.
Every polling system should undergo an assessment of its statistical accuracy. This 'confidence interval' should be made public before the election begins. For a candidate to win they must win by a statistically significant margin -- more votes than the confidence interval. If neither candidate exceeds the margin of error for the polling system used then there is a tie. What to do with a tie is another question but that does not take away the fact that there is a tie.
With the possible exception of a heated campaign for fourth grade class president, elections just aren't that tidy. The fact is that votes get lost, miscounted, etc. The reality is that in a largish election such as in the city of Seattle or the largest election, that of the President, there is no way to state with certainty what the exact number of votes for each candidate is. Ignoring for the moment the accusations of fraud just the logistics of aggregating all those votes is bound to cause some lack of exactness.
This 'lack of exactness' or 'statistical error' to use the parlance of our time gives rise to the possibility of a statistical tie even when one candidate has a nominally larger number of votes. It is a standard issue in statistical science that is resolved quite adequately by the science. I think it is time that we started including some of theory into practice.
Every polling system should undergo an assessment of its statistical accuracy. This 'confidence interval' should be made public before the election begins. For a candidate to win they must win by a statistically significant margin -- more votes than the confidence interval. If neither candidate exceeds the margin of error for the polling system used then there is a tie. What to do with a tie is another question but that does not take away the fact that there is a tie.
Thursday, June 02, 2005
Wednesday, June 01, 2005
With the Memorial Day weekend behind us I had a few moments to think about the purpose of the holiday. Between the beers and the amazing BBQ it is easy to forget that we are remembering fallen soldiers who have fought in past and current wars.
I would like to thank the soldiers who liberated my grandparents from German concentration camps. I have no idea who these soldiers are specifically but still feel grateful for their role in the my life and the lives of my entire family.
With the mass graves that are being uncovered in Iraq today I can't help but feel an affinity with those who survived the genocidal campaigns of Saddam Hussein. To me 'Never Again' means zero tolerance for genocide. I hope the US finds the courage to provide rescue to the Darfuruians although a Commentary article entitle "Can the Sudan be Saved" by Roger Sandall (December 2004) would have us believe that it is not quite possible.
I would like to thank the soldiers who liberated my grandparents from German concentration camps. I have no idea who these soldiers are specifically but still feel grateful for their role in the my life and the lives of my entire family.
With the mass graves that are being uncovered in Iraq today I can't help but feel an affinity with those who survived the genocidal campaigns of Saddam Hussein. To me 'Never Again' means zero tolerance for genocide. I hope the US finds the courage to provide rescue to the Darfuruians although a Commentary article entitle "Can the Sudan be Saved" by Roger Sandall (December 2004) would have us believe that it is not quite possible.
Subscribe to:
Posts (Atom)