psych (version 2.1.9) Plot means and confidence intervals for multiple groups


One of the many functions in R to plot means and confidence intervals. Meant mainly for demonstration purposes for showing the probabilty of replication from multiple samples. Can also be combined with such functions as boxplot to summarize distributions. Means and standard errors for each group are calculated using describeBy.

Usage,group,data=NULL, by.var=FALSE,,ylab =NULL,xlab=NULL, main=NULL, 
ylim= NULL, xlim=NULL, 
eyes=TRUE,alpha=.05,sd=FALSE,labels=NULL,v.labels=NULL,v2.labels=NULL, add.labels=NULL,
colors=c("black","blue","red"), lty,lines=TRUE, 



A data frame or matrix


A grouping variable


If using formula input, the data file must be specified


A different line for each group (default) or each variable

Is the grouping variable categorical (TRUE) or continuous (FALSE


y label


x label


title for figure


if specified, the y limits for the plot, otherwise based upon the data


if specified, the x limits for the plot, otherwise based upon the data


Should 'cats eyes' be drawn'


alpha level of confidence interval. Default is 1- alpha =95% confidence interval


sd=TRUE will plot Standard Deviations instead of standard errors


X axis label


For a bar plot legend, these are the variable labels, for a line plot, the labels of the grouping variable.


the names for the 2nd grouping variable, if there is one


if !NULL, then add the v2.labels to the left/right of the lines (add.labels="right")


where to place text: below, left, above, right


How long should the top of the error bars be?


Draw error bars for groups > min.size


add=FALSE, new plot, add=TRUE, just points and error bars


Draw a barplot with error bars rather than a simple plot of the means


Should the s.e. be corrected by the correlation with the other variables?


groups will be plotted in different colors (mod n.groups). See the note for how to make them transparent.


line type may be specified in the case of not plotting by variables


By default, when plotting different groups, connect the groups with a line of type = lty. If lines is FALSE, then do not connect the groups


Where should the legend be drawn: 0 (do not draw it), 1= lower right corner, 2 = bottom, 3 ... 8 continue clockwise, 9 is the center


The first plot symbol to use. Subsequent groups are pch + group


How many lines/inch should fill the cats eyes. If missing, non-transparent colors are used. If negative, transparent colors are used. May be a vector for different values.


if specified, the means, sd, n and perhaps of se of a data set to be plotted

other parameters to pass to the plot function e.g., lty="dashed" to draw dashed lines


Graphic output showing the means + x% confidence intervals for each group. For ci=1.96, and normal data, this will be the 95% confidence region. For ci=1, the 68% confidence region.

These confidence regions are based upon normal theory and do not take into account any skew in the variables. More accurate confidence intervals could be found by resampling.

The results of describeBy are reported invisibly.


Drawing the mean +/- a confidence interval is a frequently used function when reporting experimental results. By default, the confidence interval is 1.96 standard errors (adjusted for the t-distribution).

Improved/modified in August, 2018 to allow formula input (see examples) as well as to more properly handle multiple groups.

Following a request for better labeling of the grouping variables, the v.lab option is implemented for line graphs as well as bar graphs. Note that if using multiple grouping variables, the labels are for the variable with the most levels (which should be the first one.)

This function was originally just a wrapper for error.bars but has been written to allow groups to be organized either as the x axis or as separate lines.

If desired, a barplot with error bars can be shown. Many find this type of plot to be uninformative (e.g., ) and recommend the more standard dot plot.

Note in particular, if choosing to draw barplots, the starting value is 0.0 and setting the ylim parameter can lead to some awkward results if 0 is not included in the ylim range. Did you really mean to draw a bar plot in this case?

For up to three groups, the colors are by default "black", "blue" and "red". For more than 3 groups, they are by default rainbow colors with an alpha factor (transparency) of .5.

To make colors semitransparent, set the density to a negative number. See the last example.

See Also

See Also as error.crosses, error.bars and error.dots


Run this code
#The generic plot of variables by group SATV + SATQ ~ gender,data=sat.act)  #formula input SATV + SATQ ~ gender,data=sat.act,v.lab=cs(male,female)) #labels + SATQ ~ education + gender, data =sat.act) #see below[1:4],sat.act$gender,legend=7) #specification of variables[1:4],sat.act$gender,legend=7,labels=cs(male,female))

#a bar plot[5:6],sat.act$gender,bars=TRUE,labels=c("male","female"),
    main="SAT V and SAT Q by gender",ylim=c(0,800),colors=c("red","blue"),
    legend=5,v.labels=c("SATV","SATQ"))  #draw a barplot
#a bar plot of SAT by age -- not recommended, see the next plot + SATQ ~ education,data=sat.act,bars=TRUE,xlab="Education",
   main="95 percent confidence limits of Sat V and Sat Q", ylim=c(0,800),
   v.labels=c("SATV","SATQ"),colors=c("red","blue") )
#a better graph uses points not bars
#use formulat input
  #plot SAT V and SAT Q by education + SATQ ~ education,data=sat.act,TRUE, xlab="Education",
     main="self reported SAT scores by education",
     v.lab =c("HS","in coll", "< 16", "BA/BS", "in Grad", "Grad/Prof"))
#make the cats eyes semi-transparent by specifying a negative density + SATQ ~ education,data=sat.act, xlab="Education",
     main="self reported SAT scores by education",density=-10,
     v.lab =c("HS","in coll", "< 16", "BA/BS", "in Grad", "Grad/Prof"))

#use labels to specify the 2nd grouping variable, v.lab to specify the first  ~ education  + gender,data=sat.act, xlab="Education",
     main="self reported SAT scores by education",density=-10,
     v.lab =c("HS","in coll", "< 16", "BA/BS", "in Grad", "Grad/Prof"),

#now for a more complicated examples using 25 big 5 items scored into 5 scales
#and showing age trends by decade 
#this shows how to convert many levels of a grouping variable (age) into more manageable levels.
data(bfi)   #The Big 5 data
#first create the keys 
 keys.list <- list(Agree=c(-1,2:5),Conscientious=c(6:8,-9,-10),
        Extraversion=c(-11,-12,13:15),Neuroticism=c(16:20),Openness = c(21,-22,23,24,-25))
 keys <- make.keys(psychTools::bfi,keys.list)
 #then create the scores for those older than 10 and less than 80
 bfis <- subset(psychTools::bfi,((psychTools::bfi$age > 10) & (psychTools::bfi$age < 80)))

 scores <- scoreItems(keys,bfis,min=1,max=6) #set the right limits for item reversals
 #now draw the results by age
#specify the particular colors to use$scores,round(bfis$age/10)*10,by.var=TRUE,
      main="BFI age trends",legend=3,labels=colnames(scores$scores),
        xlab="Age",ylab="Mean item score",
        v.labels =cs(10-14,15-24,25-34,35-44,45-54,55-64,65-74))
#show transparency$scores,round(bfis$age/10)*10,by.var=TRUE,
      main="BFI age trends",legend=3,labels=colnames(scores$scores),
        xlab="Age",ylab="Mean item score", density=-10, 
        v.labels =cs(10-14,15-24,25-34,35-44,45-54,55-64,65-74))
# }

Run the code above in your browser using DataLab