>> -----Original Message----- >> From: [EMAIL PROTECTED] >> [mailto:[EMAIL PROTECTED] On Behalf Of Irene Mantzouni >> Sent: Friday, May 02, 2008 7:52 AM >> To: [EMAIL PROTECTED] >> Subject: [R] points size in plots >> >> Dear list, >> >> >> >> I would like to produce a plot of variables where the size of >> the points will be indicative of their standard errors. >> >> How is that possible? >> >> >> >> Thank you!
Do you mean (a) the larger the standard error, the bigger the point, or (b) the larger the standard error, the smaller the point? (b) makes more statistical sense. Anyways, you can use cex. An example with two variables, u and v, in the case of (a) u <- rnorm(10,2,3) v <- rnorm(10,5,4) se.u <- runif(10,1,4) se.v <- runif(10,4,7) x <- 2:11 plot(x,u,cex=se.u/max(se.u)) points(x,v,cex=se.v/max(se.v),pch=19) And in the case of (b) u <- rnorm(10,2,3) v <- rnorm(10,5,4) se.u <- runif(10,1,4) se.v <- runif(10,4,7) x <- 2:11 plot(x,u,cex=(1/se.u)/max(1/se.u)) points(x,v,cex=(1/se.v)/max(1/se.v),pch=19) If you have NAs in your variables, use the na.rm=TRUE argument of max(). If the standard errors are very different, you can multiply the quantity evaluated for cex with a positive constant, say u <- rnorm(10,2,3) v <- rnorm(10,5,4) se.u <- runif(10,1,4) se.v <- runif(10,4,7) x <- 2:11 w=2.5 plot(x,u,cex=w*(1/se.u)/max(1/se.u)) points(x,v,cex=w*(1/se.v)/max(1/se.v),pch=19) HTH Ruben ______________________________________________ R-help@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.
