Over several years, I have used microcomputers for most of my statistical analyses. This article is a summary of some of my preferences and thoughts from this experience. Some of the ideas expressed here, of course, are distillations and re-expressions of the thoughts of others.

When you use a microcomputer for your analyses, you assume all the roles traditionally associated with a computer center. You are responsible for the purchase of your machines, supplies and program selection. One problem that may arise is that you may not have a program you need or it may not exist in a version for your computer. If this is the case, you must either find a program that does the job you want or move the data set to a different computer. This problem is less severe now than it was a few years ago because many major statistical packages are available.