RMA: Software for Reduced Major Axis regression
v. 1.17 (September 3, 2004)

Andrew J. Bohonak
San Diego State University

RMA is a fast and simple application for reduced major axis regression (RMA).  RMA is more appropriate than standard ordinary least squares (OLS) regression when the independent variable x is measured with error (see Sokal and Rohlf, Biometry ).  McArdle (1988) suggests as a rule of thumb that RMA should be used when the error rate in x exceeds one-third of the error rate in y.

The RMA application estimates error for the slope and intercept of RMA regression using three methods:

  1. standard linear regression approximations (Sokal and Rohlf)
  2. jackknifing over cases
  3. bootstrapping over cases

For population genetic applications of reduced major axis regresssion, see IBDWS.

RMA Manual (pdf format)

RMA for PPC Macintosh (binhexed, self expanding achive)

RMA for Windows (self expanding achive)
NOTE for Windows users:  There is a known bug which happens with some data sets on some versions of Windows. The first analysis you do after launching RMA gives the correct answers. Subsequent analyses give the wrong answers. Because I am no longer developing the RMA program, I won't be trying to figure out the source of this compiler error. I would suggest that you use RMA for JAVA, or quit and relaunch the program after every analysis.

Java logoRMA for JAVA
courtesy Kim van der Linde, Florida State University
Bohonak, A. J. and K. van der Linde. 2004. RMA: Software for Reduced Major Axis regression for Java.


Documents labeled PDF are in Portable Document Format and require Adobe Reader to view and print. Download Adobe Reader free from the Adobe website.