[R] Hmisc, Design, summary.Design plot- changing confidence intervals, adding color or decreasing font size

[Frank E Harrell Jr]

David Winsemius wrote:
> 
> On Jul 18, 2009, at 11:14 AM, Elizabeth Stanny wrote:
> 
>>
>> Frank,
>>
>> I had already tried q=c(0.7,0.8,0.9,0.95) as an argument in 
>> plot.summary.Design and it did not work (i.e., CIs were not plotted).  
>> Could you point me to an example using plot.summary.Design that uses q 
>> as argument and has output?  I would greatly appreciate it.
> 
> I am encountering an error when using any number of probabilities other 
> than 5 and I think it relates to how the confbar col= defaults are set.
> 
> plot(s, log=TRUE, at=c(.1,.5,1,1.5,2,4,8), q=c(0.7,0.8,0.9,0.95))
> Error in confbar(nbar - (i - is + 1) + 1, effect[i], se[i], q = q, type 
> = "h",  :
>   q and col must have same length
> 
> ?confbar
> 
>  Using the example on the help page with a modified q vector of length 5 
> gives output:
> 
>  plot(s, log=TRUE, at=c(.1,.5,1,1.5,2,4,8), q = c(0.6, 0.8, 0.95, 0.975, 
> 0.99))
> 
> Experimentation shows that providing a col vector of the same length as 
> q also succeeds:
> 
> plot(s, log=TRUE, at=c(.1,.5,1,1.5,2,4,8), q=c(0.7,0.8,0.9,0.95), 
> col=gray(c(0, 0.25, 0.75, 1)))
> 

Exactly, as per the documentation.  I have just added a sentence to the 
help file to make this more clear.

Thanks
Frank

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