An opinion:
Subject: QUALITY Digest - 17 Aug 1999 to 18 Aug 1999
Date: Thu, 19 Aug 1999 00:04:38 -0400
From: Automatic digest processor <LISTSERV@PUCC.PRINCETON.EDU>
Reply-To: TQM in Manufacturing and Service Industries Discussion List <QUALITY@PUCC.BITNET>
To: Recipients of QUALITY digests <QUALITY@PUCC.PRINCETON.EDU>
There is one message totalling 77 lines in this issue.
Topics of the day:
1. DOE/Statistical Analysis Software
----------------------------------------------------------------------
Date: Wed, 18 Aug 1999 11:54:02 -0400
From: Jay Warner <a2q@execpc.com>
Subject: Re: DOE/Statistical Analysis Software
RNav@aol.com wrote:
> My department at work is in need of a software package for statistical
>
> analysis and DOEs. Does anyone have experience with any to offer a
> recommendation?
>
> Robert Drensek, CQE, CQA, CRE, CMI
> Quality Engineer
> RNAV@AOL.COM
(coming in late is what I get for working off the web for 36 hrs!)
It is no longer available commercially, but...
The single best piece of software for solving real engineering problems through honest DoE was and is Catalyst/BBN. It was designed expressly for engineers to use. The math is buried underneath, and is sound. the output is visual, informative, and easily interpretable. It makes recommendations for operating conditions for optimum response, as well as answering 'why?' It can handle up to 12 simultaneous responses. It presents planning information in an excellent format for working 'in the shop.' For clarity and simplicity of operation, it can't be beat.
Drawbacks? You can't buy it any more. It only runs on a Mac (when they wrote it, Windows couldn't do the required dynamic graphics).
I'm not aware of software available today that will match Catalyst. I'm looking at some, but no summary results yet. Most available software will function quite well for a reasonably knowledgeable analyst. Stat-Ease, JMP (from SAS) work reasonably well. Avoid any software that must be 'introduced' to you for more than 1 hr. I have a functional Excel spread sheet; DOE-KISS (an Excel add-on) is available for a nice price, and works pretty smoothly.
MiniTab is 'classical' in approach & analysis. It answers the quesiton 'why' well, but did less well on 'how should I run' questions, i.e., illustrating response surfaces near an optimum last time I looked. It has the advantage of a huge user base, esp. in business schools. Since so many people have always used it, there is a tendency to stick with it. When should we move to a new software, with new capabilities?
SAS is for professional analysts. SAS and some other routines are bascially 1-3 months and come up for air. If your designed experimetns require this level of number crunching, maybe it could be planned for easier analysis. Planning requires more intellectual effort, but is less $ 'effort' over the project life. Yes, training can have value.
Also please keep in mind, that the major impediment to using DoE (which we all should do lots of) is the thinking mode used. If you already have done 1 or more designed experiments, then software is an issue. If you haven't done one yet, or if your work hasn't given you the huge jump in information, plus production/capability/sales impact, then it's time to get started.
Jay
--
Jay Warner
Principal Scientist
Warner Consulting, Inc.
4444 North Green Bay Road
Racine, WI 53404-1216
USA
Ph: (414) 634-9100
FAX: (414) 681-1133
email: quality@a2q.com
web:
http://www.a2q.com
The A2Q Method (tm). What do you want to improve today?
Any recommendation on DOE/Statistical Analysis Software?
Linear Mode
