Box Plots

To demonstrate creating both single and multiple box plots we will use the data in the following example.

Example: The tables below list the number of police officer deaths in the United States (as of April 2012) for Northeastern and Midwestern states.

Northeastern States		Midwestern States
Maine	85	Wisconsin	354
New Hampshire	33	Michigan	550
Vermont	22	Illinois	984
Massachusetts	320	Indiana	378
Rhode Island	43	Ohio	763
Connecticut	135	Missouri	628
New York	1,327	North Dakota	50
Pennsylvania	722	South Dakota	52
New Jersey	469	Nebraska	157
		Kansas	231
		Minnesota	223
		Iowa	157

Source: www.nleomf.org

Single Box Plots

To demonstrate creating a single box plot we will use the Northeastern states data.  We'll begin by creating a list.

> northeast = c(85, 33, 22, 320, 43, 135, 1327, 722, 469)

Once we have our list created we can use the boxplot command to create the box plot.

> boxplot(northeast)

By default, the boxplot command separates outliers from the rest of the data, identifying them with an open circle.  We can add the argument range = 0 if we do not want outliers separated.

> boxplot(northeast, range = 0)

We can also add the argument main to title our plot.

> boxplot(northeast, main = "Police Officer Deaths in the Northeastern United States")

We can also output the numerical information from our graph using the plot = FALSE argument.

> boxplot(northeast, plot = FALSE)
$stats
     [,1]
[1,]   22
[2,]   43
[3,]  135
[4,]  469
[5,]  722

$n
[1] 9

$conf
       [,1]
[1,] -89.36
[2,] 359.36

$out
[1] 1327

$group
[1] 1

$names
[1] "1"

The $stats section tells us

• [1] the minimum usual value
• [2] the first quartile
• [3] the median/second quartile
• [4] the third quartile
• [5] the maximum usual value

The $out section identifies the values of any outliers.

Multiple Box Plots

To demonstrate creating a multiple box plot we will use both data sets in the problem above.  We'll begin by creating a list for the Midwestern data

> midwest = c(354, 550, 984, 378, 763, 628, 50, 52, 157, 231, 223, 157)

Once we have our lists created we can use the boxplot command to create the multiple box plot.

> boxplot(northeast, midwest)

As with single box plots we can use the main argument to title our plot.  We can also use the names argument to label the individual box plots.  In this case we will create a list containing the names "Northeastern United States" and "Midwestern United States."

> boxplot(northeast, midwest, main = "Police Officer Deaths in the United States", names = c("Northeastern United States", "Midwestern United States"))