Re: Benford's Law

[Kromkowski@aol.com]

With regard to Steve's response:

I think one of the assumptions is that the data make up a "collection". So I 
don't think your c-chart data would be helpful. Because, those data aren't 
normally added up to get some other results, as is the case in the figures 
entered on a tax return or the figures added up to get to net income.  For an 
example, you might take a look at the election data from Florida by county. I 
looked at it before all the recounts (fit pretty well), but not after. Who 
knows?  There is also the question of number of members of a Benford set. 
(E.g., the Florida election results may be said to have 67 members by county, 
many more by precinct.)  As I understand it, even 67 is rather small.  (I 
wonder if a p-chart is good predictive tool for the expectable variation for 
each digit?)

I think that the quick and dirty article I posted may not have done Benford's 
law justice. Unfortunately, I am too busy to publish additional sources, 
right now.

John David Kromkowski
Kromkowski@aol.com  Telephone:  (410) 377-6248 Facsimile:  (410) 372-0624