Bruce,

Thanks for the encouragement re: doing it yourself. I've got some other leads from this thread.

Larry

>>I spoke to SPSS (a big stats company) and they will license their components for distribution, but at a whopping $20,000.00 startup fee plus royalty on each sale.
>
>And SAS will cost similar...those are the only two stat products I use. Hopefully, someone will provide another vendor for you. But in case you get stuck:
>
>This is easier said than done, but you can build your own statistical code in vfp for redistribution without too much difficulty, as a rule. Vfp is excellent for statistical work - it's not the fastest number-crunching product in the world by any means (C++ is notably faster), but having native Rushmore makes coding a dream, compared to something like C++. And building your own is not just nice due to cost-savings, but because you can customize it much easier in a vfp app.
>
>Multiple regression is not all that hard to do, given the basic formula. I've done it in the past in C or Pascal as an assignment, I forget which language now.
>
>A medium-level stats book will have both the formula and some examples to help.
>
>(This idea is just in case you can't find anything else)