Adding bidirectional error bars to points on scatter plot in ggplot

Question

I am trying to add x and y axis error bars to each individual point in a scatter plot. Each point represents a standardized mean value for fitness for males and females (n=33).

I have found the geom_errorbar and geom_errorbarh functions and this example

ggplot2 : Adding two errorbars to each point in scatterplot

However my issue is that I want to specify the standard error for each point (which I have already calculated) from another column in my dataset which looks like this below

 line     MaleBL1   FemaleBL1  BL1MaleSE BL1FemaleSE
     3  0.05343516  0.05615977 0.28666600   0.3142001
     4 -0.53321642 -0.27279609 0.23929438   0.1350793
     5 -0.25853484 -0.08283566 0.25904025   0.2984323
     6 -1.11250479  0.03299387 0.23553281   0.2786233
     7 -0.14784506  0.28781883 0.27872358   0.2657080
    10  0.38168220  0.89476555 0.25620796   0.3108585
    11  0.24466921  0.14419021 0.27386482   0.3322349
    12 -0.06119015  1.42294820 0.32903199   0.3632367
    14  0.38957538  1.66850680 0.30362671   0.4437925
    15  0.05784842 -0.12453429 0.32319116   0.3372879
    18  0.71964923 -0.28669563 0.16336556   0.1911489
    23  0.03191843  0.13955703 0.34522310   0.1872229
    28 -0.04598340 -0.35156017 0.27001451   0.1822967

'line' is the population (n=10 individuals in each) from where each value comes from my x,y variables are 'MaleBL1' & 'FemaleBL1' and the standard error for each populations for males and females respectively 'BL1MaleSE' & 'BL1FemaleSE'

So far code wise I have

p<-ggplot(BL1ggplot, aes(x=MaleBL1, y=FemaleBL1)) +
geom_point(shape=1) +    
geom_smooth(method=lm)+ # add regression line
xmin<-(MaleBL1-BL1MaleSE)
xmax<-(MaleBL1+BL1MaleSE)
ymin<-(FemaleBL1-BL1FemaleSE)
ymax<-(FemaleBL1+BL1FemaleSE)    
geom_errorbarh(aes(xmin=xmin,xmax=xmax))+
geom_errorbar(aes(ymin=ymin,ymax=ymax))

I think the last two lines are wrong with specifying the limits of the error bars. I just don't know how to tell R where to take the SE values for each point from the columns BL1MaleSE and BL1FemaleSE

Any tips greatly appreciated

Roland Roland · Accepted Answer · 2014-05-06T15:09:02

You really should study some tutorials. You haven't understood ggplot2 syntax.

BL1ggplot <- read.table(text=" line     MaleBL1   FemaleBL1  BL1MaleSE BL1FemaleSE
     3  0.05343516  0.05615977 0.28666600   0.3142001
     4 -0.53321642 -0.27279609 0.23929438   0.1350793
     5 -0.25853484 -0.08283566 0.25904025   0.2984323
     6 -1.11250479  0.03299387 0.23553281   0.2786233
     7 -0.14784506  0.28781883 0.27872358   0.2657080
    10  0.38168220  0.89476555 0.25620796   0.3108585
    11  0.24466921  0.14419021 0.27386482   0.3322349
    12 -0.06119015  1.42294820 0.32903199   0.3632367
    14  0.38957538  1.66850680 0.30362671   0.4437925
    15  0.05784842 -0.12453429 0.32319116   0.3372879
    18  0.71964923 -0.28669563 0.16336556   0.1911489
    23  0.03191843  0.13955703 0.34522310   0.1872229
    28 -0.04598340 -0.35156017 0.27001451   0.1822967", header=TRUE)

library(ggplot2)
p<-ggplot(BL1ggplot, aes(x=MaleBL1, y=FemaleBL1)) +
  geom_point(shape=1) +    
  geom_smooth(method=lm)+ 
  geom_errorbarh(aes(xmin=MaleBL1-BL1MaleSE,
                   xmax=MaleBL1+BL1MaleSE),
                 height=0.2)+
  geom_errorbar(aes(ymin=FemaleBL1-BL1FemaleSE,
                    ymax=FemaleBL1+BL1FemaleSE),
                width=0.2)

print(p)

enter image description here

Btw., looking at the errorbars you should probably use Deming regression or Total Least Squares instead of OLS regression.

Adding bidirectional error bars to points on scatter plot in ggplot

1 Answers