[R] Plot means with error bars - A novice needs help

Lawrence Hanser larry at hanser.net
Thu Oct 9 14:22:47 CEST 2008

Dear Colleagues,

I think the allure of bargraph.CI is that it makes it easy to plot
standard errors--though you still have to fiddle with the options to
get the size of confidence interval you want. And there is the
(strong) dislike of dynamite plots by some (the ink to info argument).

Box and whisker plots have such a tradition of meaning in the boxes
and whiskers that to add standard errors requires some additional work
and explanation, and even then standard error markings are easily lost
among the usual box and whisker plot markings.

I have looked only briefly at tplot(), but don't recall seeing an easy
way to add standard errors. Have I missed it?  Would it be an easy
add?  A bad idea?



On 10/8/08, hadley wickham <h.wickham at gmail.com> wrote:
> On Tue, Oct 7, 2008 at 11:31 PM, Michael Just <mgjust at gmail.com> wrote:
>> Hello,
>> I'd appreciate a suggestion on how to construct plots (barplots?) that use
>> means on the Y axis instead of density/count. I'd also like to use groups
>> and plot error or confidence interval bars on these graphs. I know this is
>> a
>> read the manual situation. I'd appreciate help with what to read, or your
>> benevolence with some sample code.
> Here's an alternative suggestion - don't use bars, use dots.  Bar
> plots with standard errors are sometime called dynamite plots
> (probably because they should be blown up).  See
> http://biostat.mc.vanderbilt.edu/twiki/bin/view/Main/DynamitePlots and
> http://emdbolker.wikidot.com/blog:dynamite for some reasons not to use
> them and possible alternatives.
> I hope that package authors who provide methods to make these plots
> easy will reconsider.
> Hadley
> --
> http://had.co.nz/
> ______________________________________________
> 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.

Sent from my mobile device

More information about the R-help mailing list