I've been doing more and more work that requires developing and validating models (computing the Maximum Likelihood Estimators, running goodness-of-fit tests, etc). I'm likely to be doing even more of this stuff in the future. Any of you have experience with statistical packages that take care of this stuff (SAS, SPSS, R, etc.)? Any of them really good? Any of them really bad?

I know that SAS is damn near the industry standard in aerospace, but I’ve not worked with it much.

R’s good, though it can be persnickety about the formatting of the data it will accept.

This may help

This is Mary, Dan’s wife – I’ve used the less expensive software packages JMP and Minitab. JMP in my opinion (Six Sigma Blackbelt) is trash. JMP doesn’t call any test by it’s actual statistical name. So if you want to do a t-test (how simple) it takes forever to find the right menu option. Minitab on the other hand uses the correct names for everything – and has good graphical output – especially useful if you ever need to explain to a statistical novice. I never could find out how to do a f-test in JMP so I wound up doing it in Excel.

SAS is JMP’s big, too expensive, brother. I’ve heard it is more powerful, but not nearly as many use it due to its price. However, if its little brother is any indication, it may not be worth it.

Dan again…

Hopefully that was helpful, if you have any more detailed questions for her just email me at yuehrmah@hotmail.com