Date:         Sat, 7 Feb 2004 07:23:01 GMT
Reply-To:     Arthur Tabachneck <art297@NETSCAPE.NET>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Arthur Tabachneck <art297@NETSCAPE.NET>
Subject:      Re: SPSS vs. SAS, redux

Bill,

As another statistician and programmer (of sorts), as well as a researcher
and manager, in the hope that one of those influential little birdies from
SAS happens to be following this thread, I couldn't agree more.

I've been using SAS since 1974, and SPSS since 1975, thus have seen
virtually all of the flavors which both have offered over the years.

When I need to accomplish a task quickly, and don't have a template handy
containing a sample of the desired code, I find SPSS to be the clear winner.
I also proclaim it the winner when it comes to working with labels, handling
missing values, producing graphs, and providing reviewable and modifiable
source code.

Yes, SAS clearly wins when it comes to allowing the greatest array of
statistical controls, complex data manipulations, and file handling
capabilities but, in my opinion, they have failed miserably when it comes to
ease of use and, interestingly (in their newest features) user quality
control capabilities.  In that last regard, I was shocked to discover that
Enterprise Miner actually "had" a nice, easy to use graphical interface, but
wouldn't provide the user with the needed source to do such things as modify
the code for repeated (though possibly modified) use or to review the code
to ensure that it really was doing everything precisely as intended.

The one issue I disagree with is the one concerning putting strong
statistical tools into the hands of non-statisticians.  What it has
accomplished more, I think, is having put good statistical tools out of the
reach of many otherwise good statisticians and researchers.  Many excellent
SAS programmers have virtually no statistical background, can accomplish
virtually any task that the system can perform, but may need some guidance
when it comes to using SAS statistical procedures.  That doesn't make them
either bad programmers, or necessarilly evil statistician wannabees, just
people who don't know a lot about statistics and, hopefully, are aware of
their limitations.

I am quite aware of my own programming limitations and have never hesitated
to bring in a more qualified SAS consultant whenever needed.

Art
-----------
"Thompson Bill T Contr USAFSAM/FEC" <Bill.Thompson@BROOKS.AF.MIL> wrote in
message
news:E3D75F0AC6CA2545A301F27ADECD562E02F57467@casino.brooks.af.mil...
> I agree that you need to be a "statistician" and a "programmer (of
sorts)",
> however let me express my 2 cents.
>
> I am both a statistician and a programmer who uses both SPSS and SAS.  I
> find that the point and click method in SPSS sometimes helps save time.
For
> example, while I do code in both SPSS and SAS when using SPSS if I am
doing
> a simple analysis, I can run the point and click method and have SPSS
> generate the code.  I can then take that code, review it to ensure it is
> doing what I want, copy it and past to a "syntax" file and then alter the
> code in any way I see fit.  This methodology, at times, saves time in
> writing code.  Since as a statistician I hopefully "know" what I am doing,
I
> can protect against some of the errors that ocurr in just using the point
> and click method and still save some time.
>
> I do have concerns about those who know enough statistics to make them
> dangerous (opps, may be stepping on my own toes) using the point and click
> methodology without the "knowledge base" required to understand what they
> have produce or interpret the results accurately.
>
> Just my 2cents.
>
> Billy
>
> -----Original Message-----
> From: Jay Weedon [mailto:jweedon@EARTHLINK.NET]
> Sent: Thursday, February 05, 2004 2:27 PM
> To: SAS-L@LISTSERV.UGA.EDU
> Subject: Re: SPSS vs. SAS, redux
>
>
> On 5 Feb 04 17:54:58 GMT, david.mcnulty@QUESTINTL.COM wrote:
>
> >My knee jerk reaction to any statistical analysis delivered by a point
and
> >click interface is to enquire about
> >1) the statistical qualifications of the person doing the analysis
> >2) the number of years they have used the software package
> >
> >SPSS have made a deliberate attempt to make statistics point and click.
> >Virtually anyone can get a statistical analysis out of SPSS but do the
> >results mean anything?
> >
> >I am not knocking SPSS,  I just value the protection offered by a command
> >line interface and a modicum of "Technical know-how needed to operate".
>
> I guess a valid response would be: To use SPSS you need to be a
> statistician. To use SAS you need to be a statistician and a
> programmer.
>
> JW