R help - Dec 2009 - multiple plots using summary in rms package

Dear All,

I wonder if someone can point me in the right direction here.  I'm working
with the rms library, R 2.9.2 under Windows XP.

I'm trying to arrange two plots side by side for a colleague.  mfrow or
mfcol do not seem to work, however, so I am obviously missing something
important.  I know that there have been changes in the graphics from Design
to rms, but am just not sure where to find specific documentation about this
particular issue.

Below is the code I'm using.  If I run the code except for the last plot
call, the first plot is correctly produces in the first column with empty
space for the second plot in col 2.  But when I add the second plot call, it
just overwrites the first plot in col1 on the left.

If I run the code below using something like plot(x~y) and plot(a~b) instead
of the two rms summary objects, it works fine, so I am assuming there is
something about rms I am missing.

I'd be happy to provide some data if that would help.

Any guidance would be greatly appreciated.

Thanks,

Mike Babyak
Duke University Medical Center


########################################################
#summary of dietary variables broken out by Groups A, B, and C
sumfc<-summary(Group~fcgrain+fcveg+fcfruit+fcmeat+fcdairy+fcnsl+
        fcfat+fcsatfat+fcsweet+fcsod,
method='reverse', overall=F,  test=F)

sumpc<-summary(Group~pcgrain+pcveg+pcfruit+pcmeat+pcdairy+
        pcnsl+pcfat+pcsatfat+pcsweet+pcsod,
    method='reverse', overall=F,  test=F)

par(mfcol=c(1,2),oma=c(1,0,4,0))
plot(sumfc, which='categorical',
main='Full Compliance',pch=c('A','B','C'))
 Key(0,-.1)

plot(sumpc, which='categorical',
main='At Least Partial
Compliance',pch=c('A','B','C'))
mtitle("Dietary Adherence")

	[[alternative HTML version deleted]]

Mike,

You are not using rms functions in your example; you are using Hmisc's 
summary.formula function and its plot method.  For method='reverse', 
multiple plots may be produced by the plot method so par(mfrow=c( )) is 
not well defined.  There is also a problem with the dotchart2 function 
(called by plot.summary.formula.reverse) in this context, as it wants to 
call plot.new().

Frank

Mike Babyak wrote:
> Dear All,
> 
> I wonder if someone can point me in the right direction here.  I'm
working
> with the rms library, R 2.9.2 under Windows XP.
> 
> I'm trying to arrange two plots side by side for a colleague.  mfrow or
> mfcol do not seem to work, however, so I am obviously missing something
> important.  I know that there have been changes in the graphics from Design
> to rms, but am just not sure where to find specific documentation about
this
> particular issue.
> 
> Below is the code I'm using.  If I run the code except for the last
plot
> call, the first plot is correctly produces in the first column with empty
> space for the second plot in col 2.  But when I add the second plot call,
it
> just overwrites the first plot in col1 on the left.
> 
> If I run the code below using something like plot(x~y) and plot(a~b)
instead
> of the two rms summary objects, it works fine, so I am assuming there is
> something about rms I am missing.
> 
> I'd be happy to provide some data if that would help.
> 
> Any guidance would be greatly appreciated.
> 
> Thanks,
> 
> Mike Babyak
> Duke University Medical Center
> 
> 
> ########################################################
> #summary of dietary variables broken out by Groups A, B, and C
> sumfc<-summary(Group~fcgrain+fcveg+fcfruit+fcmeat+fcdairy+fcnsl+
>         fcfat+fcsatfat+fcsweet+fcsod,
> method='reverse', overall=F,  test=F)
> 
> sumpc<-summary(Group~pcgrain+pcveg+pcfruit+pcmeat+pcdairy+
>         pcnsl+pcfat+pcsatfat+pcsweet+pcsod,
>     method='reverse', overall=F,  test=F)
> 
> par(mfcol=c(1,2),oma=c(1,0,4,0))
> plot(sumfc, which='categorical',
> main='Full Compliance',pch=c('A','B','C'))
>  Key(0,-.1)
> 
> plot(sumpc, which='categorical',
> main='At Least Partial
Compliance',pch=c('A','B','C'))
> mtitle("Dietary Adherence")

-- 
Frank E Harrell Jr   Professor and Chair           School of Medicine
                      Department of Biostatistics   Vanderbilt University