Re: st: Stata vs SPSS

[Richard Williams <Richard.A.Williams.5@ND.edu>]

At 03:47 PM 10/13/2006, Raphael Fraser wrote:
> I am being persecuted by some of my students for having to learn Stata
> rather than SPSS. "SPSS is so easy to use" they say. "Why do we have
> to type every thing?" one student complained. This is always the case
> in my Intro to Stata classes but at the end of the course they show
> much appreciation for learning Stata especially when they have to do
> data manipulation. But this year is different. There is a student
> rebellion. They see me as a "Stata Dictator" and is desirous of
> "Democracy." So I have decided to be more "diplomatic" with a Stata vs
> SPSS comparison.
Interesting to hear some of the comments that have come out.  I 
myself have never had to produce 102 esoterically formatted tables; 
for me outreg2 and estout are more than adequate for my needs, 
although obviously this is a problem for others.  I've always 
suspected that SPSS and SAS devote 90% of their code to user 
interface and formatting output, and I can see why based on some of 
these comments.

However, I suspect that isn't the kind of thing your students were 
complaining about!  Who are these people, what is the field of study, 
what sorts of things are they going to need to do in their future 
careers?  If all they are going to do is run frequencies and OLS 
regressions with non-weighted data sets, it may not matter too much 
which package they are going to use.

You could, of course, be like me, and teach both packages!  That is 
more historical accident than anything else, as I have used SPSS for 
almost 30 years and only started using Stata a few years ago.  But, 
it doesn't hurt to be multilingual.  I don't think either package is 
all that difficult, at least for basic stuff.

Some of the things I stress to students are
1) Much easier hypothesis testing, e.g. testing that the effects of 
x1 and x2 are equal is much easier in Stata
2) Stata is much faster for typing basic commands - can anybody 
remember SPSS's syntax for regression and logistic regression?  Sure, 
the SPSS pulldown menus make things easier but they also slow you up.
3) Just try doing something like robust standard errors in SPSS.  A 
student asked me how to do that once, and we finally found about a 10 
page SPSS program that did it!
4) Complicated survey designs can be handled easily by Stata.  Not so 
with SPSS.
5) Try running a logistic regression with a large data set in both 
programs.  Stata is far faster.
6) Various advanced methods are implemented or better implemented in 
Stata.  For example, categorical data analysis.
7) SPSS documentation is free but is little more than a syntax 
guide.  The Stata manuals are expensive but vastly more 
informative.  Not only do you get the syntax, but you get a basic 
overview of the statistical theory behind the commands.

Again, though, it is not that hard to know both SPSS and Stata.  SAS 
may be more of a challenge.  I think the main difficulties come when 
you are doing complicated data manipulations or highly customized 
outputting of results.  But I'm not sure how many people need to do 
that on a routine basis.


-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
FAX:    (574)288-4373
HOME:   (574)289-5227
EMAIL:  Richard.A.Williams.5@ND.Edu
WWW (personal):    http://www.nd.edu/~rwilliam
WWW (department):    http://www.nd.edu/~soc 
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/