Robert J. MacG. Dawson <[EMAIL PROTECTED]> wrote:


: dennis roberts wrote:
:> 
:>   i would say that what bob has indicated below is BIAS ... not consistency

:       Of course it's bias... but bias *does* imply inconsistency. It's a
: simple implication between two important concepts, so makes a good point
: to mention.

:       -Robert Dawson

what??? bias does NOT imply consistency. here is an estimator that is biased
but consistent: xhat = sum(x_i)/(n+a) where i goes from 1 to n and a is any
arbitrary constant. if Ex(x_i) = m (the population mean), then Ex(xhat) ~= m
(so it is biased) but xhat is consistent because a/n goes to zero (so it is
asymptotically unbiased) and its variance goes to zero.

biasedness and consistency are independent.

g.


=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
                  http://jse.stat.ncsu.edu/
=================================================================

Reply via email to