Statalist The Stata Listserver

[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: Stata vs SPSS

From   "Earnhart, Benjamin J" <>
To   <>
Subject   RE: st: Stata vs SPSS
Date   Mon, 16 Oct 2006 13:03:26 -0500

No, generally not.  Predictive Analytics is a very expensive package,
marketed towards business; your everyday user at a University is not
going to cough up $3000 or whatever for a bunch of tools s/he will never
use, though I can imagine some Business schools using it.

Most of the point-and-click stuff does the normal type things you do in
most packages (regressions of various types, ANOVA, recoding, selecting
cases and the like).  For stats classes and substantive analyses in the
Social Sciences and the like, SPSS is pretty much the same as Stata in
many ways (though many others have pointed out the differences, and
remember that one of them is that with SPSS you pay for different
subsets of capabilities unlike Stata where you get an all-inclusive
package you can add features to for free).

For Predictive Analytics, models we're used to like OLS, Multinomial,
factor analysis and the like can be *part* of such a system, but it
usually goes far beyond that, with things like weighting by the
desirability of (or damage caused by) an outcome.  Much more data-driven
than theory-driven, it may even involve neural nets and other
machine-learning techniques.  The ideal is to throw a bunch of data
against the wall, see what sticks, and let the computer make a decision
for you.  If it works, can be great, but done poorly, can really stink.

Hope that helps?

*Ben Earnhart
*Computer Consultant and 
*ICPSR Representative
*Department of Sociology and 
*College of Liberal Arts
*University of Iowa
*(319) 335-2887

> -----Original Message-----
> From: 
> [] On Behalf Of Nick Cox
> Sent: Monday, October 16, 2006 12:41 PM
> To:
> Subject: RE: st: Stata vs SPSS
> So, is this what you can get in SPSS by pointing 
> and clicking? 
> Nick 
> Earnhart, Benjamin J
> > Data mining, in the sense of using an AI to fit a model.  
> > Done well, it
> > can be very powerful (think of all the credit card and 
> insurance fraud
> > such systems prevent), but done poorly (for example training and
> > predicting on the same dataset), there are some serious drawbacks.
> Nick Cox
> > > From the same horse's mouth: 
> > > 
> > > "SPSS is recognized as a leader in the 
> > > predictive analytics market space. 
> > > Predictive analytics, which combines 
> > > advanced analytics and decision optimization, 
> > > will continue to be a focus for the organization 
> > > as it seeks to increase marketplace understanding 
> > > of the business benefits that predictive 
> > > analytics provides." 
> > > 
> > > Anyone know what this means? 
> > > >
> *
> *   For searches and help try:
> *
> *
> *

*   For searches and help try:

© Copyright 1996–2021 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index