Date: Fri, 29 Oct 2010 15:43:29 +0200
Reply-To: Joakim Englund <joakim.englund@GMAIL.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Joakim Englund <joakim.englund@GMAIL.COM>
Subject: Re: Proc gchart - plotting bars + standard deviations
In-Reply-To: <5826426ECF51494E9F2BAD2C3F3E7310212108CF@MAIL.dhw.state.id.us>
Content-Type: text/plain; charset=ISO-8859-1
Thanks for your input Chris! It does seam like an annotate dataset might
well alleviate my problem. However, I can't get it to work properly, which I
think is due to that I have group= and subgroup= as options under the vbar
statement. My proc gchart code:
proc gchart data=PDPKall3b (where=(biomat="&biomat"));
vbar trtnum / subgroup=trtnum group=daynum nozero frame /*type=mean clm=90
ERRORBAR=both*/ anno=anno raxis=axis1
maxis=axis2 gaxis=axis3 coutline=black woutline=*1* sumvar=pdmean
patternid=subgroup legend=legend2;
run;
I've calculated the means before hand, so I don't use the type=mean option.
Do you know how to edit the code for the anno dataset provided in your link,
to get it to work properly for my gchart code?
Thanks!
/Joakim
2010/10/28 Murphy, Christopher F. - CO 1st <MurphyC@dhw.idaho.gov>
> Joakim,
>
> There is a way to create your graph in SAS, but you need to use a data
> step to create an annotate dataset that tells SAS to draw custom error
> bars. See Sample 24883 (and code) on the SAS/GRAPH Samples Output
> Gallery page:
>
> http://support.sas.com/sassamples/graphgallery/PROC_GCHART_Graph_Element
> s_Statistical_Graph_Enhancements.html
>
> http://tinyurl.com/35t55vn
>
> Chris
>
>
>
> -----Original Message-----
> From: Joakim Englund [mailto:joakim.englund@GMAIL.COM]
> Sent: Thursday, October 28, 2010 8:55 AM
> Subject: Proc gchart - plotting bars + standard deviations
>
> Hi,
>
> I'm using proc gchart to plot some bars on top of which I would like to
> add "T"s indicating standard deviations. Option ERRORBAR= under the VBAR
> statement allegedly allows you to do this. However, the idea here is
> that you specify confidence limits (using the CLM= option) to plot
> rather than standard deviations. If I had only one bar, I could
> backwards calculate what significance level to specify in the CLM=
> option to yield the length of one standard deviation in the plot (as
> suggested in SAS help for this situation).But I have several bars,
> implying that I would then need to somehow respecify CLM= for every bar,
> which I don't think can be done (or I don't know how to do it).
> Furthermore, I use the subgroup= option (since I need to plot different
> treatments for different days in the same plot, i.e.
> DAY is the subgroup) and SAS says the errorbar= option is not allowed
> using the subgroup statement.
>
> Is there no way to create a graph in SAS that meets my simple
> requirements??
>
> KR,
> Joakim
> The information contained in this email may be privileged, confidential or
> otherwise protected from disclosure. All persons are advised that they may
> face penalties under state and federal law for sharing this information with
> unauthorized individuals. If you received this email in error, please reply
> to the sender that you have received this information in error. Also,
> please delete this email after replying to the sender.
>
| 
