R help - Feb 2013 - a priori power analysis for glm, family = poisson

Dear R/statistics wizards,



This may be more of a statistics question than an R one, but I’m hoping
that someone has the time to help.  I’ve already consulted a few local
statisticians and I’ve not yet received a clear answer.  Before I start an
experiment I want to conduct an a priori power analysis test for a
Generalized Linear Model, family = poisson.  The response variables will be
counts.  I want to look at the effects of 5 temperatures and 3 salinities,
and their interaction in 15 different combinations (5 temperatures each
testing 3 different salinity levels).   I am hoping to learn the power I
would achieve if I used a sample size of 18 in each of the combinations,
for a total sample size N = 270.



Does such a power analysis exist?  Can I just use a power analysis for a
linear model since the glm is an extension of that (which seems to make
sense), or maybe even one from a 2 way anova? Can someone show me the path
of righteousness in R, or point me in the right direction please?



Thank you very much for your valuable time.

Chad

	[[alternative HTML version deleted]]

Probably the best way of doing this is via simulation. Have a look at
chapter 5 of:

Bolker, B. 2008. Ecological models and data in R. Princeton Univ. Press,
Princeton, NJ.

for an introduction and a bunch of examples.

Joerg


On Wed, Feb 20, 2013 at 5:49 AM, Chad Widmer <cw647 at st-andrews.ac.uk>
wrote:
> Dear R/statistics wizards,
>
>
>
> This may be more of a statistics question than an R one, but I?m hoping
> that someone has the time to help.  I?ve already consulted a few local
> statisticians and I?ve not yet received a clear answer.  Before I start an
> experiment I want to conduct an a priori power analysis test for a
> Generalized Linear Model, family = poisson.  The response variables will be
> counts.  I want to look at the effects of 5 temperatures and 3 salinities,
> and their interaction in 15 different combinations (5 temperatures each
> testing 3 different salinity levels).   I am hoping to learn the power I
> would achieve if I used a sample size of 18 in each of the combinations,
> for a total sample size N = 270.
>
>
>
> Does such a power analysis exist?  Can I just use a power analysis for a
> linear model since the glm is an extension of that (which seems to make
> sense), or maybe even one from a 2 way anova? Can someone show me the path
> of righteousness in R, or point me in the right direction please?
>
>
>
> Thank you very much for your valuable time.
>
> Chad
>
>         [[alternative HTML version deleted]]
>
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>