[Nipy-devel] rm 2-way ANOVA

josef.pktd@gmai... josef.pktd@gmai...
Thu Nov 12 20:06:36 CST 2009


On Thu, Nov 12, 2009 at 8:41 PM, Christian Brodbeck
<christianmbrodbeck@gmail.com> wrote:
>
> Is this a random effects model? If so, is it balanced so the "usual" ANOVA
> tables can be used?
>
> I guess seeing the R code of what you are fitting would be helpful.
>
> I am using the R aov function:
>
> aov(data~(factor1*factor2)+Error(subject/(factor1*factor2)))
>
> (after "Two-way Within Subjects ANOVA"
> on http://www.personality-project.org/r/r.anova.html)
> My data are event related potentials for 4 conditions (2x2) for 22 subjects
> (each subject provides an ERP for each condition).
>
> My rpy2 code is the dependent_2way function in
> http://bazaar.launchpad.net/~christianmbrodbeck/eelbrain/new/annotate/head:/rpy2stats.py .
>
> Christian
> _______________________________________________
> Nipy-devel mailing list
> Nipy-devel@neuroimaging.scipy.org
> http://mail.scipy.org/mailman/listinfo/nipy-devel
>
>

There are all the pieces for doing anova with ols and dummy variables
in statsmodels, but I never finished cleaning up.

http://bazaar.launchpad.net/~scipystats/statsmodels/trunk/annotate/head%3A/scikits/statsmodels/sandbox/regression/try_ols_anova.py

only oneway anova is verified against the NIST data, (oneway also has
a direct non ols implementation)

http://bazaar.launchpad.net/~scipystats/statsmodels/trunk/annotate/head%3A/scikits/statsmodels/sandbox/regression/anova_nistcertified.py

Josef



More information about the Nipy-devel mailing list