Graphing Data with R

Chapter 14. The Bland-Altman Plot

Note

If you are not interested in programming, you can skip this chapter with no loss of continuity.

Assessing Measurement Reliability

This chapter illustrates the flexibility of R. You cannot find the type of graph examined here in base R, and it can take a bit of effort to find such a specialized kind of plot among the thousands of available packages. I decided to write my own function to accomplish the task and have included it here. Since I wrote this function, I discovered that at least two packages include this plot; I’ll introduce one of them here. There is a little more typing to do in this chapter than in most others, with a dataset to enter in a spreadsheet or text file and a relatively long R function to type. There is an alternate, shorter version of the function at the end of the chapter that you can use instead if you like.

The Bland-Altman plot is a tool used to assess the agreement between two measurement techniques, or the reliability/repeatability of a measurement. It is also known as the “Tukey mean-difference plot.”

Bland and Altman (1986) give the data listed in Table 14-1 on measuring peak expiratory flow rate (PEFR), in liters/minute, with two different types of meters, a Wright flow meter and a Mini Wright flow meter. Two measurements were made with each meter, on each subject. The object here is to determine whether the Mini meter gives substantially the same readings as the Wright meter and could, therefore, be substituted ...