Measure of dispersion for geometric mean

[D.R. Wahlgren]

Gang,
I hope you can help with a general stats question that is only
tangentially related to SPSS.  We have some variables that are
positively skewed and for which we are running parametric analyses on
log-transformed versions.  For descriptive purposes, we are
presenting geometric means (calculated by taking the antilog of the
mean of the transformed variable).

The problem is that we need a measure a dispersion to complement the
geometric mean, but I've been told that we cannot similarly take the
antilog of the SD of the transformed variable.

Is there a way to present dispersion for a skewed variable that makes
sense when presented with a geometric mean?  Second, is there a way
to generate this in SPSS or would it be a manual calculation similar
to how we derive the geometric mean?

thanks,
Dennis
--
Dennis R. Wahlgren, M.A.

Center for Behavioral Epidemiology and Community Health
San Diego State University
http://www.cbeach.org

"Poets say science takes away from the beauty of the stars--mere
globs of gas atoms.  Nothing is 'mere.'  I too can see the stars on a
desert night, and feel them.  But do I see less or more?"
--Richard Feynman

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Marta Garcia-Granero]

D.R. Wahlgren wrote:

> I hope you can help with a general stats question that is only
> tangentially related to SPSS.  We have some variables that are
> positively skewed and for which we are running parametric analyses on
> log-transformed versions.  For descriptive purposes, we are
> presenting geometric means (calculated by taking the antilog of the
> mean of the transformed variable).
>
> The problem is that we need a measure a dispersion to complement the
> geometric mean, but I've been told that we cannot similarly take the
> antilog of the SD of the transformed variable.

That's correct.
>
> Is there a way to present dispersion for a skewed variable that makes
> sense when presented with a geometric mean?
You can use 95%CI for the GM

> Second, is there a way
> to generate this in SPSS or would it be a manual calculation similar
> to how we derive the geometric mean?
Look for a previous message of mine (title: "Re: 95% conf. intervals for
geometric means?", date: February 8)

HTH,
Marta García-Granero

>
> thanks,
> Dennis
> --
> Dennis R. Wahlgren, M.A.
>
> Center for Behavioral Epidemiology and Community Health
> San Diego State University
> http://www.cbeach.org
>
> "Poets say science takes away from the beauty of the stars--mere
> globs of gas atoms.  Nothing is 'mere.'  I too can see the stars on a
> desert night, and feel them.  But do I see less or more?"
> --Richard Feynman
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD
>

--
For miscellaneous statistical stuff, visit:
http://gjyp.nl/marta/

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD