specify se for boxplot?

[AC Del Re]

Hi All,

Is there a way to adjust the standard errors (se) that are used to derive the error bars using geom_boxplot() ? I am trying to create boxplots for categorical moderator levels using meta-analytic procedures and it seems the stats used to draw these plots are not ideal for this method (?). If someone has advice on how I could adjust the se then it might work.

Thanks,

AC

[hadley wickham]

Hi AC,

What do you mean standard errors? There aren't any in a boxplot.
Maybe you want to compute your own summary statistics and display them
with geom_crossbar.

Hadley

> --
> You received this message because you are subscribed to the ggplot2 mailing
> list.
> To post to this group, send email to ggp...@googlegroups.com
> To unsubscribe from this group, send email to
> ggplot2+u...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/ggplot2

--
Assistant Professor / Dobelman Family Junior Chair
Department of Statistics / Rice University
http://had.co.nz/

[AC Del Re]

Oh, right, the boxplot wouldnt be appropriate in this case. I will look at geom_crossbar. Thanks, Hadley

AC