On Thu, 10 Jul 2003, Russell Barbour wrote:
> To measure auto-coorelation Moran's I is compared to the expected "I" that
> would be generated from a random distribution. My question is if the observed
> Moran's I is more than one standard deviation higher than expected "I" could
> that be consi
To measure auto-coorelation Moran's I is compared to the expected "I" that
would be generated from a random distribution. My question is if the observed
Moran's I is more than one standard deviation higher than expected "I" could
that be considered significant auto-correlation?
Thanks
Russell