Adding error bars to xyplot()

Hi,

I want to add error bars to a plot generated with xyplot.  I've  
tried both errbar() and plotCI(), but in both cases the points are  
not in the same place.  It's as if the two functions are using a  
different frame of reference for the plotting area.

for example:
means <- c(92.5, 92.25, 90.9, 91.0, 94.15, 90.05) #means
time <- c(1,1,2,2,3,3) #occasion variable
group <- rep(c("n","u"),3) #grouping variable
SE <- c(2.22, 1.66, 2.10, 1.43, 2.31, 1.57) #standard errors

#Plot the graph
xyplot(means ~ time, groups=group, type="o", lwd=3, main="Change  
Over Time by Condition", scale=list(cex=2), xlab="Test Number",  
auto.key=T)

#first attempt
require(Hmisc)
errbar(x = time, y = means, yplus = means + SE, yminus = means - SE,  
add=T, col=c("blue", "hotpink"))

#second attempt, same result
require(gplots)
plotCI(x = time, y = means, uiw = SE, add=T)

Thanks in advance!

- Jon

-- 
Jon Zadra
Department of Psychology
University of Virginia
P.O. Box 400400
Charlottesville VA 22904
(434) 982-4744
email: zadra at virginia.edu
<http://www.google.com/calendar/embed?src=jzadra%40gmail.com>

On Apr 21, 2010, at 6:24 PM, Jon Zadra wrote:

Hi,

I want to add error bars to a plot generated with xyplot.  I've tried 
both errbar() and plotCI(), but in both cases the points are not in 
the same place.  It's as if the two functions are using a different 
frame of reference for the plotting area.

Yes. plotCI uses base graphics while xyplot is a lattice function.

for example:
means <- c(92.5, 92.25, 90.9, 91.0, 94.15, 90.05) #means
time <- c(1,1,2,2,3,3) #occasion variable
group <- rep(c("n","u"),3) #grouping variable
SE <- c(2.22, 1.66, 2.10, 1.43, 2.31, 1.57) #standard errors

#Plot the graph
xyplot(means ~ time, groups=group, type="o", lwd=3, main="Change Over 
Time by Condition", scale=list(cex=2), xlab="Test Number", auto.key=T)

#first attempt
require(Hmisc)
errbar(x = time, y = means, yplus = means + SE, yminus = means - SE, 
add=T, col=c("blue", "hotpink"))

A base graphics function.

#second attempt, same result
require(gplots)
plotCI(x = time, y = means, uiw = SE, add=T)

Another base graphics function.

Since you already have Hmisc why not check out Harrell's xYplot which 
provides CI's in a grid function:
From the xYplot examples:

require(Hmisc)
dfr <- expand.grid(month=1:12, continent=c('Europe','USA'),
 sex=c('female','male')) set.seed(1)
dfr <- upData(dfr, y=month/10 + 1*(sex=='female') +
                          2*(continent=='Europe') +
                          runif(48,-.15,.15),
                   lower=y - runif(48,.05,.15),
                   upper=y + runif(48,.05,.15))
xYplot(Cbind(y,lower,upper) ~ month|continent, 
subset=sex=='male',data=dfr)

Thanks in advance!

- Jon

-- 
Jon Zadra
Department of Psychology
University of Virginia
P.O. Box 400400
Charlottesville VA 22904
(434) 982-4744
email: zadra at virginia.edu
<http://www.google.com/calendar/embed?src=jzadra%40gmail.com>

David Winsemius, MD
West Hartford, CT

I took a look at xYplot() but it is beyond over-complex, at least  
for my level, and the documentation isn't too clear.

Is there a simple way to make functions like plotCI() and errbar()  
use the lattice reference rather than base graphics?

I took a look at xYplot() but it is beyond over-complex, at least for my
level, and the documentation isn't too clear.

Is there a simple way to make functions like plotCI() and errbar() use the
lattice reference rather than base graphics?

It seems like creating a plot of two lines across three time points
separated into two groups should be a very basic, simple thing for a
statistics program to do. ?Is there really no way to simply pass a model to
a plotting function and get something that includes error bars? ?I just
wonder if I"m missing how to do this "the easy way."