At 1:18 PM -0500 23/4/01, Jon Cryer wrote:
>These examples come the closest I have seen to having a known variance.
>However, often measuring instruments, such as micrometers, quote their
>accuracy as a percentage of the size of the measurement. Thus, if you
>don't know the mean you also don't know the variance.
Certainly many measurements do have errors that are best given as a
percent of the reading. In such cases, the error usually is a
"constant" percent, not a constant absolute amount. To put it
another way, the log of the readings has a normally distributed error
that is independent of the reading. So you should perform all your
analyses on the log-transformed variable, and express all your
outcomes as percent differences or changes. Otherwise your analyses
are riddled with non-uniform error (heteroscedasticity).
Will
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================