Stats packages

K. Lee Koetzner KK2543A at american.edu
Wed Mar 8 13:15:39 EST 1995


I had always heard (secondhand or thirdhand) that Statistica gave the most reli
able numerical results for your analyses this side of mainframe BMDP, so when I
 started having trouble replicating results in JMP (macintosh), I switched over
 to Statistica for Windows (we had a license for CSS/Dos, so it was fairly chea
p). Short version: I like Statistica. A lot. It's easy to use, fast, and has ye
t to give a result I find unlikely. There are things I don't like about it (no
32-bit or PowerPC versions until the end of the year, maybe; the transfer betwe
en different types of analyses feels wierd), but it has a lot to recommend it--
including a very powerful looking scripting language. I haven't worked with it
yet, but it looks easier than writing SAS/VMS batch files, yet just as powerful
.
I'd be the last person to argue stats with Wilkinson, but without knowing more
about the alleged code theft, I'm not letting it keep me away.



More information about the Neur-sci mailing list