Adding Error Bars in R with ggplot

Question

I would like to make a plot in R plotting two categorical variables (treat) by LWP.

I would like to change the color of the dots (the fill) to correspond with the variable rootpatH (one dot in each treatment).

I would also like to add a notation if the differences are significant or not..

Plants.means<-aggregate(Plants[10:44],Plants[c(6,9)],mean,na.rm=TRUE)

SE<-function(x)(sd(x,na.rm=TRUE)/ (length(x)-sum(is.na(x)))^0.5)


ggplot(Plants.means, aes(x=Treat, y=Yield, group=1)) +
geom_errorbar(width=.1, aes(ymin=Yield-SE(Plants.means$Yield), ymax=Yield+SE(Plants.means$Yield))) +
geom_point(shape=21, size=3, fill="white") +
ylim(0,1750)

I also want to change the error bars to represent 95% CI instead of SE.

Treat is categorical rootpatH is categorical LWP_meas2 is continuous

enter image description here

`SE` is a function, not a numeric value; presumably you want `SE(something)` ... — Ben Bolker, Feb 02 '13 at 23:16
how would I add text "ns" and "*" next to each set of points to denote if they are significantly different or not? — user1977802, Feb 03 '13 at 03:47

score 1 · Answer 1 · answered Feb 03 '13 at 01:19

As Ben says you probably want something like this (of necessity not tested since no data was offered:

with( Plants.means,
               errbar(x =Treat, y = LWP_Meas2, 
                                yplus =  LWP_Meas2 + SE( LWP_Meas2), 
                                yminus = LWP_Meas2 - SE( LWP_Meas2), 
       add=T,na.rm=TRUE, col=c("blue", "hotpink"))
      )

score 0 · Accepted Answer · answered Feb 03 '13 at 03:44

error <- function(x)(qt(0.975,df=length(x)-1)*sd(x)/sqrt(length(x)))

ggplot(Plants.means,aes(x=Treat, y=Yield, colour=rootpatH, group=rootpatH)) +
geom_errorbar(aes(ymin=Yield-error(Plants.means$Yield),ymax=Yield+error(Plants.means$Yield)), colour="black",width=.1, position=pd) +
geom_point(position=pd, size=4)

Adding Error Bars in R with ggplot

2 Answers2