Skip to content

dnorm

3 messages · Sibylle Stöckli, Jim Lemon, Greg Snow

Original
Sibylle Stöckli
# 
Dear R-users

Idea:
Plot a dnorm line using specific mean/sd to complete a histogram (skewed). xs:range of y-values, ys: dnorm function

Problem:
I expected to multiply the ys function with the sample size (n=250-300). I was wondering about a factor between 12'000 and 30'000 to match the size of the dnorm line with the specific histogram.

Thanks
Sibylle

hist(Biotree[Ld,]$Height2008, main="Larix decidua", ylim=c(0,50), xlab="Tree Height 2008 (cm)",col="aquamarine", font.main=3, cex.axis=0.8)
xs<-0:650
ys<-dnorm(xs, mean=397.8, sd=97.6)
lines(xs,ys*12000)
Jim Lemon
#
On 09/23/2010 05:42 PM, "Sibylle St?ckli" wrote:
Hi Sibylle,
You can use one of the "rescale" functions (one is in the plotrix 
package) like this:


Ld.hist<-hist(Biotree[Ld,]$Height2008, main="Larix decidua",
  ylim=c(0,50), xlab="Tree Height 2008 (cm)",
  col="aquamarine", font.main=3,cex.axis=0.8)
xs<-0:650
ys<-dnorm(xs, mean=397.8, sd=97.6)
lines(xs,rescale(ys,0:max(Ld.hist)))

Jim
Greg Snow
# 
You need to take into account the bin width as well (hence the extra multiple you asked about), but it is simpler to just include prob=TRUE in the hist call, then you do not need to do any adjustment on the y-values of the reference distribution.