heath@gpu.utcs.toronto.edu (Todd Heatherton) (05/20/88)
Does anyone know of a powerful statistics program for the Mac, which also conforms to the Mac interface. I know that Statview 512+ is a great program in terms of ease of use, but it does not do complex ANOVA designs or post-hoc tests. As well, it does not have the capacity for anything beyond 2-way contingency tables. I have seen CLR ANOVA from Clear Lake Research, but I do not like the way they perform their statistical tests (unweighted means). I had seen an old version of SYSTAT, but it seemed far from user-friendly. The program I am looking for in the Mac world would be similar to PowerStat in the PC world; a program which is powerful but easy to use. Any suggestions or comments?
6029334@pucc.Princeton.EDU (Robert G. Trevor) (05/20/88)
In article <1988May19.182607.21352@gpu.utcs.toronto.edu>, heath@gpu.utcs.toronto.edu (Todd Heatherton) writes: > > >Does anyone know of a powerful statistics program for the Mac, which >also conforms to the Mac interface. I know that Statview 512+ is a I have Beta version of RATS for the Macintosh. Although a little buggy, especially if you want to use the built in compiler, its a very powerful and useful package. They are working hard to remove remaining bugs - you can get beta now with free upgrade to commercially relaesed version, or wait for finished version. Tech support is good (they sent me new beta when I found bugs in previous version), and the built-in procedures are very robust. RATS has a good name amongst economists/econometricians, and is available in mainframe, IBM-PC versions as well as source licence for M68020 work-station systems. Main draw-back commercially is the user-interface (it shows its bloodline); then that shouldn't worry people who are using high-powered stats routines anyway. Facilities are mainly for linear models (although has some non-linear routines and they are adding more); they include time series (time and frequency domain), forecasting (static/dynamic, deterministic/stochastic), multiple regression, 2SLS, 3SLS, SUR, NLLS, logit, probit, Kalman filter, a (presently buggy) structured programming language (including matrix language) and much more. (+ BBS for new procedures) If this sounds like a rave - it is. I make my living from using such packages. RATS is one of the best. Only GAUSS beats it, but (SIGH) thats not yet available for the MAC. (They claim the're about to hire a programmer to start porting it, and that it will be available in the Fall - pigs can fly too.) Contact: VAR Econometrics PO Box 1818 Evanston, IL 60204-1818 Ph: (312) 864-8772 Rob Trevor MaBell: (609) 452-4051 Bitnet: 6029334@PUCC UseNet: 6029334@PUCC.Princeton.Edu
whiteheada@byuvax.bitnet (05/21/88)
DataDesk Professional from Odesta should hit the streets next week. It has some excellent features that you are asking. Call them and they will send a brochure on it. Armand Whitehead BYU
g451252772ea@deneb.ucdavis.edu (0040;0000009857;0;327;142;) (05/21/88)
The current version of Systat is impressive. Still only 'user-cordial' to use Wilkenson's own phrase, but much improved over 2.x (up to 3.1 now). I also use StatView and like its interface much better. Systat is a citable, professional tool. Yes I've earned a few dollars reporting bugs (one stat, one trivial) in it, but I liked the rapid feedback Wilkenson gave. For anova, there's some non-commercial stuff- GANOVA, from UCLA, comes to mind. The MGLH program in systat is still limited by it's memory allocation to small designs when unequal-n repeated measures & such are set up. Sigh- the regression approximation used in GANOVA works much better in those cases. good luck. Ron Goldthwaite / UC Davis, Psychology and Animal Behavior 'Economics is a branch of ethics, pretending to be a science; ethology is a science, pretending relevance to ethics.'