Stata has enormous advantages over SPSS in just about every area with the
possible exception of traditional ANOVA designs. SAS is the most
comprehensive package and widely used as a standard, but Stata is rapidly
developing. SAS now claims that SAS never stood for Statistical Analysis
System and much of its focus is in areas other than statistical analysis.
However, they maintain an outstanding staff and do a good catching up with
statistical developments. Stata makes no claims to be other than a
statistics package and all of its energies go to this development. Stata has
1000s of user written programs that can be installed on the fly when they
are needed. These may vary in quality but I've not seen any serious problems
with them. Stata really had most of its development after the internet and
is far more internet aware than SPSS or SAS. The command 'update all' will
update your Stata version to the latest issue and this seems to happen at
least monthly. This often involves new procedures and improvements on old
procedures. The command findit xxxx will produce a list of relevant user
written programs that can be installed by clicking a button.

A check of the "movie" on logistic regression at
http://www.ats.ucla.edu/stat/stata/seminars/stata_logistic/ illustrates how
Stata and various user supplied programs let you do far more with logistic
regression than SPSS or SAS. The same site has movies for SAS and SPSS and a
comparison of these movies shows where the action is.

Stata lacks a truly introductory manual like SAS or SPSS, but for anybody
with experience with another package, Stata is easy to learn. The
documentation and help system on Stata are the clearest I've seen.

I'm not an  epidemiologist, but Stata's interest in development in this area
is evident in an entire collection of specialized procedures for
epidemiology.

The pricing of Stata and SAS is far more user friendly than SPSS. When you
buy Stata Intercool or Stata SE you get the entire thing. With the grad plan
the university pricing is $129 and $259, respectively. This is a purchase,
not a one year lease. They sell a 4 volume reference manual and a larger
user's guide on the grad plan for $129. SPSS sells bits and pieces. Version
12 of SPSS can finally deal with complex sample designs but only for a
couple applications (means, crosstabs) and this is an extra cost module.
SPSS is very marketing oriented and this may be why they divide their
package into a bunch of pieces knowing you need all the pieces and need to
pay a fortune when you add them all up.

It is true that a SAS expert can make a good living simply as a SAS expert
and not too many can do this with Stata or SPSS.

Alan Acock
----- Original Message ----- 
From: "Silvert, Henry" <[EMAIL PROTECTED]>
To: "mac55" <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>
Sent: Friday, March 26, 2004 5:28 AM
Subject: Re: spss v. stata


| I worked on an HIV/AIDS project in the past and we used SAS. My feeling is
that SAS allows for more the use of more macros and is the preferred
package. If you become proficient in SAS I think that you might also be more
marketable to people in the pharmaceutical industry.
|
|
|
| Henry M. Silvert  Ph.D.
| Research Statistician
| Management Excellence
| The Conference Board
| 845 3rd Avenue
| New York, NY 10022
| Tel. No.: (212) 339-0438
| Fax No.: (212) 836-3825
| www.conference-board.org
|
|  -----Original Message-----
| From: mac55 [mailto:[EMAIL PROTECTED]
| Sent: Thursday, March 25, 2004 5:05 PM
| To: [EMAIL PROTECTED]
| Subject: spss v. stata
|
| Within the end of this year, I will be starting an epidemiological PhD
| disseration.  I have used SPSS for years.  Most of my colleagues and
| my students who I have to help use SPSS also.  However, in my last two
| epi classes they have been using stata.  It looks more
| straightforward.
|
| My question is what are the advantages and disadvantages.  I know that
| it mostly depends on peference.  But, I was thinking about giving
| stata a try.  It looks a bit easier (they are both easier than SAS
| from what I hear).
|
| Thanks!
| Marc
| .
| .
| =================================================================
| Instructions for joining and leaving this list, remarks about the
| problem of INAPPROPRIATE MESSAGES, and archives are available at:
| .                  http://jse.stat.ncsu.edu/                    .
| =================================================================
|
|
| .
| .
| =================================================================
| Instructions for joining and leaving this list, remarks about the
| problem of INAPPROPRIATE MESSAGES, and archives are available at:
| .                  http://jse.stat.ncsu.edu/                    .
| =================================================================

.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to