Hello,

Hi Baron,

My bad, you are correct. my mistake was that i was remembering vmr as

the ratio of the stddev (σ) squared to the mean, with std dev (σ)

being the variance squared. so vmr is indeed the variance / mean

ratio, as the name implies. my stats class memories are a little bit

rusty.

By the way, i meant to say Coefficient of variation, not dispersion.

http://en.wikipedia.org/wiki/Coefficient_of_variation
http://en.wikipedia.org/wiki/Index_of_dispersion (also called variance

to mean).

Thanks for clearing that up, sorry for wasting some time ...

Cheers,

Romain.

On 11 mar, 13:20, Baron Schwartz <

ba...@xaprb.com> wrote:

> Romain,

>

> > * variance V(X) is the mean of the squared distances to the mean

> > * std dev is the squared root of the variance

> > * variance to mean = the ratio of the variance (σ) squared to the mean

> > (µ) : σ^2/ μ

>

> Why do you say variance-to-mean is the variance *squared* to the mean?

> That would be variance-squared-to-mean. Is this a standard

> terminology that is used by statisticians?

>

> > * coefficient of dispersion = variance to mean : σ / µ

>

> I am not familiar with the coefficient of dispersion, but it sounds

> like the same thing we are calling the variance-to-mean ratio. I'm

> not a mathematician or statistician, so it doesn't surprise me that

> I'm ignorant of this. The reason we added variance-to-mean to the

> profile was because of a paper by Robyn Sands available on the Method

> R website. I did not research existing literature on it beyond that.