Analysis of Variance Explained
(and a tool to do it!)

I. Scott MacKenzie

Dept. of Computer Science and Engineering

York University

Toronto, Ontario, Canada M3J 1P3

Last update: 16-May-18

An analysis of variance, or ANOVA, is the main statistical tool used in human-computer interaction experiments. It is used to answer research questions such as "Does the time to complete a task depend on which of several interaction methods is used?" In studying the analysis of variance, I found it useful to write a program to perform the test. It was quite a grind, but I learned from the exercise. The program is in Java. While working on it, I wrote the API to explain the details.

The first version, called Anova2, dates to 2002. Recently, the app has been updated substantially. The latest version, called Anova3GUI, is GUI style and supports 12 experiment designs.

Both the program and API are used in my teachings, and they are usually placed on a web site for students to download. It seemed worthwhile to add them to the Research Notes on my home page, so here they are:

         API (read here for all the details)

         Download site (The GUI version of Anova3GUI is in GoStats.jar I recommend you use that instead of the older command-line version)

Further details, discussion, and examples are found in Chapter 6 (Hypothesis Testing) in my recent book, Human-Computer Interaction: An Empirical Research Perspective.

If you find this useful or have other comments, please get in touch.