CLICK HERE FOR BLOGGER TEMPLATES AND MYSPACE LAYOUTS »

16 January 2008

What the hell is a Keith Number?

Watching Keith Olbermann tonight, I heard a term that I have seen and/or heard over the past couple of weeks. It has to do with statistical analysis. A Keith Number is, to quote from its Wikipedia entry, "...a Keith number or repfigit number (short for repetitive Fibonacci-like digit) is an integer N>9 that appears as a term in a linear recurrence relation with initial terms based on its own digits."

Clear as mud????

And that was the clearest explanation I was able to find. So, my question is this: If the clearest explanation requires, at best, a minor in advanced mathematics from an accredited university, is it really necessary or appropriately applicable to include that term with the statistical analysis of the Presidential primaries currently being held?

2 Messages of Encouragement Received Today:

Jonathan Sadowski said...

The Keith Number, as he uses it, is the % of undecided voters plus the margin of error for the poll.

So if a margin of error for a poll is +/-3%, and the number of undecided voters in the poll is 10%, the Keith number for that poll would be 13%

Simon, Inc. said...

The "Keith Number" as Olbermann uses it is his own polite homage to the real Keith Number (named after Mike Keith) you found referenced on Wikipedia (which i can't understand a word of either).
Olbermann's "Keith Number" is simply the margin of error plus the percentage of undecided voters in the respective poll. It helps indicate a "real" lead than the too often "perceived" leads that most polls tout.