**=========================================================================
****Date:** Tue, 18 Jul 2006 14:12:53 -0400
**Reply-To:** Joseph Teitelman temp2 <cmtemp2@pgahq.com>
**Sender:** "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
**From:** Joseph Teitelman temp2 <cmtemp2@pgahq.com>
**Subject:** Re: SV: For those in need of running SAS using an SPSSfile
**Content-Type:** text/plain; charset=ISO-8859-1
You still haven't answered my question as to why most statisticians perfer SAS to SPSS; or why it's become common knowledge that SAS outperforms SPSS in so many respects. Have you read the pdf document at UCLA? It was written in 2005 * hardly outdated.

Moreover, you stated that you received a Ph.D. in Economics, and then in parentheses placed the term econometrics. I was unaware that
any universities award degrees specifically in econometrics. Are you trying to say that you earned a degree in economics and specialized in econometrics.

By the way, from what University have you earned your Ph.D. and at what universities have you taught. How were you able to join a statistics department as a tenure track faculty member with a degree in economics, despite your apparent expertise as an econometrician.

Personally, I received my Ph.Ds from Duke and UNC-CH. Check out US News and World Report. You'll discover that with respect to my fields of study, each of these schools is consistently ranked among the best in my areas of study.

>>> "Peck, Jon" <peck@spss.com> 7/18/2006 1:55 PM >>>
As a matter of fact, I do not agree with you in the slightest when you consider the power of the open source modules that can be plugged in to SPSS via programmability and combined with the power and elegance of the Python language and the SPSS engine. Maybe you should try it.

And as another matter of fact, I happen to have a Ph. D. in Economics (econometrics), and taught in a top-tier university in the Economics and Statistics departments for 13 years before joining SPSS.

I have no interest in badmouthing SAS or other competing products. Each has strengths and weaknesses. Instead, I try to learn what those are.

-Jon Peck
SPSS

-----Original Message-----
From: Joseph Teitelman temp2 [mailto:cmtemp2@pgahq.com]
Sent: Tuesday, July 18, 2006 12:30 PM
To: SPSSX-L@LISTSERV.UGA.EDU; Peck, Jon
Subject: Re: SV: For those in need of running SAS using an SPSSfile

I'm sure that you'd agree with me that it's patently obvious that the SPSS programming language hardly compares to SAS' SAS/IML module. There's simply no such comparison. SAS's matrix programming language is far superior to SPSS' matrix programming language.

And the question still remains: why do most statisticians choose SAS over SPSS? You tell me why such is the case. Have you been a statistics student in a graduate level program * or a mathematics student?

>>> "Peck, Jon" <peck@spss.com> 7/18/2006 12:37 PM >>>
First, let's correct the facts. SPSS in fact does have a matrix language built in. It has 18 statement types, 59 functions, and 20 operators. Users on this list have posted extensive programs using it.

Second, using the programmability features of SPSS 14, you have access to a vast array of scientifically oriented modules from third parties. For example, scipy and numpy can be downloaded free and used within SPSS.

Here is the summary description of scipy

SciPy is an open source library of scientific tools for Python. SciPy gathers a variety of high level science and engineering modules together as a single package. SciPy provides modules for statistics, optimization, integration, linear algebra, Fourier transforms, signal and image processing, genetic algorithms, ODE solvers, special functions, and more. SciPy requires and supplements NumPy, which provides a multidimensional array object and other basic functionality.

Here are a few random examples of simple things you can do with this library within BEGIN PROGRAM in SPSS. Since you can read the SPSS cases and output in this mode, the inputs can be anything in SPSS.

factorial and combination functions:

import scipy
scipy.factorial(4)
-> array(24.0)
int(scipy.factorial(4))
-> 24
scipy.factorial(4.1)
-> array(27.931753738368371) (Gamma function)
scipy.factorial(50, exact=1)
-> 30414093201713378043612608166064768844377641568960512000000000000L

matrix operations:
import scipy
A = scipy.mat('[1 3 5;2 5 1;2 3 8]')
print A
-> matrix [[1 3 5]
[2 5 1]
[2 3 8]]
print A.I
-> matrix([[-1.48, 0.36, 0.88],
[ 0.56, 0.08, -0.36],
[ 0.16, -0.12, 0.04]])
(A * A.I = identity matrix)
scipy.linalg.det(A)
-> -25

solving linear equations (nonlinear also available):
from scipy import *
A= mat('[1 3 5;2 5 1;2 3 8]')
b = mat('[10;8;3]')
Solve linear equations Ax = b...
A.I*b
or
linalg.solve(A,b)

Regards,
Jon Peck
SPSS

-----Original Message-----
From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of Joseph Teitelman temp2
Sent: Tuesday, July 18, 2006 9:15 AM
To: SPSSX-L@LISTSERV.UGA.EDU
Subject: Re: [SPSSX-L] SV: For those in need of running SAS using an SPSS file
[snip]

Next, Stat/IML is a matrix programming language which comes along with SAS. It is extremely powerful. SPSS has no matrix programming language. And from what I've been told, neither does Stata.
[>>>Peck, Jon] [snip]

Those were my impressions.