On 10/4/06, Michael Plenty <[log in to unmask]> wrote:
> My name is Mike and I work as an intern for a major consulting firm in
> Washington,DC.
>
> My company uses SPSS, as most undergraduate programs and companies, however,
> now that I'm beginning the thinking process for graduate school, most
> schools, in particular Northwestern, UChicago, and Yale, all recommend I
> become familiar with Stata and R.
>
> Are there any major differences between the three programs?
>
> If so, what? and what resources are out there to help me adjust?
While all three programs provide (to some degree) a graphical user
interface, Stata and R use tends to be more command-driven - while
SPSS has an underlying command language, it is rarely taught to
undergraduates these days. So that's the big adjustment you will
face, although Stata now has added pretty menus for the most common
statistical operations and there are add-on packages for R (like R
Commander) that also provide a GUI.
More stuff is "canned" in Stata and SPSS than in R. Take a logistic
regression (logit) model for example: in Stata and SPSS, there is a
discrete command that runs logistic regression. R expects you to know
that logit is a type of generalized linear model... so to estimate a
logit in R, you'd use glm(y ~ x, family=binomial). And, in R you
don't get p-values or t/z ratios without asking for them... so you'd
have to type summary(glm(y ~ x, family=binomial)). In Stata, you'd
just type logit y x, and you'd be done. In SPSS, you'd have to pay
for the logistic regression module (in Models?), but once you did
that... you'd point and click.
I think the main difference is in terms of "mindset"; I typically
spend more time trying to get R to do what I want than I did in Stata,
mostly because R is not a "stats package" but first and foremost a
general programming language with better-than-average support for
vector and matrix operations with a lot of stats convenience
functions--it's like if C or Perl or Python or Java came with a really
awesome stats library out of the box. Out of the bits and pieces
floating around on the Internet, you can build a DIY political science
methods toolbox in R, it's just a bit more work, although pulling in
Zelig and the stuff it depends on will get you about 90% there.
This does have one advantage: R is infinitely more flexible and is
radically easier to extend with a mix of R and compiled code. And
another key advantage over SPSS and Stata is that R's absolutely free
of charge, and you can start learning it right now - just go to
www.r-project.org and download it. There are a number of excellent
free tutorials available online.
I think in the end I have to agree with Jim Battista - you'll probably
have to learn Stata and R at some point. I don't think it's critical
to come into graduate school with that background, as you learn these
programs by *using* them, not by studying a manual or textbook, but it
could give you a leg up on your colleagues (then again, if their
undergrad methods training was as woeful as many students', just
knowing SPSS would give you a leg up).
Chris
--
Christopher N. Lawrence <[log in to unmask]>
Assistant Professor of Political Science (non-tenure-track)
Saint Louis University
109 Fitzgerald Hall
3500 Lindell Boulevard
St. Louis, Missouri 63103-1021
Website: http://www.cnlawrence.com/
**********************************************************
Political Methodology E-Mail List
Editor: Karen Long Jusko <[log in to unmask]>
**********************************************************
Send messages to [log in to unmask]
To join the list, cancel your subscription, or modify
your subscription settings visit:
http://polmeth.wustl.edu/polmeth.php
**********************************************************
|
