Just a very short comment. Point-and-click procedures can be automatically
pasted with SPSS into syntax files (in addition to the fact that one can
write the syntax directly), which avoids loosing track of past procedures,
changes and bright ideas and results. It's true that profs don't use to
teach SPSS with syntax anymore, but that is only our (profs) fault.
Personally, for undergraduates when I teach basic methodology I use SPSS, as
this is the package more usually available for non-university professionals
in Spain, and it's fairly easy for them to understand and start using. I
always emphasize using the syntax files as a complete log of EVERYTHING you
do with your data files (similar to do-files in Stata). If students learn
this trick early on there should be no huge regrets for using SPSS for the
most simple things (descriptive stats, bivariate analyses, etc.). I agree
that for more sophisticated analyses (more common in graduate school and
while doing a statistically sophisticated dissertation), Stata and other
specialised packages are much more helpful.
best,
Laura Morales
-------------------------------------------------------------------------
Dr. Laura Morales-Diez de Ulzurrun
Assistant Professor
& Coordinator of the LOCALMULTIDEM project (FP6 STREP)
Department of Political Science
University of Murcia, Spain
Ronda de Levante 10
30008 Murcia, Spain
Tel: +34-968 36 72 61
Fax: +34-968 39 83 91
www.um.es/dp-politica-administracion/curriculum/laura-eng.php
www.um.es/localmultidem
----- Original Message -----
From: "Jim Battista" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Friday, October 06, 2006 9:50 AM
Subject: Re: [POLMETH] R vs. Stata vs. SPSS
> Hey, I didn't mean to stomp on you, and I'm sure you're actually
> right-thinking and not muddleheaded. This is the same advice I
> generally offer grad students. I end up wording it strongly primarily
> because I have been That Guy that lost an interesting result from
> point-and-click stacks when I was in grad school, so I offer the advice
> in terms that are strong enough that I hope even I would have paid some
> attention to when I was a strong-willed grad student. Sorry if I gave
> offense.
>
> Walt Borges wrote:
>>
>> While Jim and I agree (apparently) that R is the best program to learn,
>
> Actually, my point was that it doesn't really matter which is the best
> to learn. To get through a program and dissertation as a
> normal-to-quantish Americanist, you can expect that you're going to have
> to learn some Stata and some R (or some boutique software) no matter
> what you use to run OLS, completely irrespective of their actual worth
> over a career-long period.
>
>> I'll even defend "point and click," -- to a point. It's helpful
>> especially when you use it to set up graphics. I really don't see value
>> in spending a whole lot of time programming the graphics. Grad students
>> don't have a lot of time, and frankly, the graphic artists are better
>> with the graphs and tables.
>
> Sure, and with Stata it's not too bad because all the point and click
> does is generate a command to pass to the CLI, so you can copy the
> command to a text file if you remember to do so.
>
> But using point and click interfaces to statistical software, at least
> for analysis, really is leading yourself into temptation. It's easy to
> acquire bad habits that then come back and bite at inopportune moments.
> It also makes it exceedingly, annoyingly difficult to help students
> deal with their actual data problems because there's no program or
> command stack to look at and see just what part of their analysis went
> blooey.
>
>> Finally, a lot of the secondary data sets we students use must be
>> converted into CSV format for use in R. Sometimes its easy to do,
>> sometimes it's not. Just depends on how much I can afford to spend this
>> semester on programs to do the conversion. Comprendo?
>
> If you have the usual full range of statistical software, the usual
> answer would be to load it into whatever it's native to and then export
> it to a csv. Even SPSS and SAS seem to be able to do this, though SAS
> probably wants to charge you $38000 for the relevant package.
>
> --
> James S. Coleman Battista
> Dept. of Political Science, Univ. of North Texas
> [log in to unmask] (940)565-4960
>
> **********************************************************
> 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
>
> **********************************************************
**********************************************************
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
**********************************************************
|
