Minitab, My Bad.

Back in We Got Mail (Part 2) I stated that we do not have a copy of Minitab. Legendary stat consultant Michael Mitchell sent me an email to let me know that, yes indeed, we do have a copy of Minitab 14 on our server. I was wrong, my bad. So, I decided to sit down and try it out for myself.

It uses pull down menus and dialog boxes which remind me a bit of SPSS. The small examples that I tried ran very quickly. I didn't read the documentation (I'm a Mac user) but I explored the menus and tried as many things as I could. I did notice that Minitab has very strong quality control procedures, but that it was lacking in areas like poisson and negative binomial regression, instrumental variable (2SLS) models and mixed models with random slopes and intercepts.

I am not a very big fan of point-and-click interfaces for stat programs. I know that Minitab can be run using syntax or commands because it is possible to run it in batch mode. However, when running interactively it is not readily apparent how to make it respond directly to commands.

So, while Minitab is a very nice program, I don't think that it would meet the needs of many of the researchers on campus. It does look like it would be a good program for students to use in thier coursework. It isn't used much at ucla because faculty teach using the stat software they use in their own reaserch. I know it really isn't fair to judge a program based on such a quick examination but Minitab doesn't seem to add anything to the "big" three (SAS, SPSS and Stata) that we already use.

pbe