[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Num Rat
//--- forwarded letter -------------------------------------------------------
> MIME-Version: 1.0
> Date: Tue, 11 Jul 95 13:22:33 -0600
> From: "zinc" <[email protected]>
> To: [email protected]
> Cc: [email protected]
> Subject: Re: Num Rat
Pat Finerty <[email protected]> wrote:
> On Tue, 11 Jul 1995 [email protected] wrote:
>
> > John Young posted:
> >
> > > He's Got Their Number: Scholar Uses Math to Foil Financial
> > > Fraud
> > >
[snip]
> > I'd bet, that
> > 5(five), 8(eight), & 9(nine) are significantly more represented across
> > the board in prices (& thus in amounts for checks & tax write offs) than
> > than their random distribution by Benford's Law or more well known tests
> > for randomness would suggest.
[snip]
>
> check amounts will also include any relevant sales tax thus skewing the
> distribution in some fashion.
>
>
> patrick finerty = [email protected] = [email protected]
> U of Utah biochem grad student in the Bass lab - zinc fingers + dsRNA!
[snip]
Yes, and some vendors will be in state (sales tax) & some out of state (no
sales tax). Furthermore, if the vendor is in state but in a different locale,
there will probably be some difference in sales tax rates as rates within
states are usually based on vendor location. Also some types of purchases
for some types of businesses/organizations/entities have various sales tax
exemptions or surcharges, again all of which varies by state & locality.
ALL of these factors will skew the distribution, eg. sales tax is usualy
*.00% or *.25%, *.50%, *.75% etc. so a cursory look shows that 0 & 5 will
be over represented due to this factor.
tjh