Hi Roland and folks:
Roland's and Susanna's comments were very interesting. I think we need
to keep in mind that unlike SPSS, R is more than a statistical package. It's
a tool. SPSS does indeed make life easy, particularly for data preprocessing (or
data "cleaning"), something that is perhaps not advisable using R (the
current versions). Also, I think it's a myth that R has a steep learning
curve. Actually, the learning curve of R is much flatter compared to SPSS or
SAS, may be because R is extremely flexible.
My personal experience, (for whatever it's worth), from no concept of R, I
learned it in about 6 hours using John Maindonald's tutorial+R-archives with
two windows opened, one for documentation and one for the IDE, and I had a
dataset ready. In comparison, to reach the same level of user comfort, it took
me as a graduate student at least three semesters work at a US university to
learn SPSS. SPSS looks deceptively easy in its point-and-click environment.
That said, I think co-teaching R and point-and-click SPSS to statistics
newbies-intermediates who are also budding social scientists can make the lives
of these social science majors a lot harder than they already are. As a result,
you'll lose a lot of people who could have converted to R, but now will
perceive R as a difficult to learn language. Either stick with SPSS, or use SPSS
for data preprocessing, and once they are up and comfortable with basic
functionalities of R, switch them to R. But not both at the same time or in
quick succession.
If I were to teach a course where SPSS and R would coexist, I'd introduce R
slowly. I'd start with SPSS, show them how to use SPSS for routine data
management tasks and OLAP etc..., and teach the basics using SPSS, and encourage
the students to write syntaxes in SPSS instead of point and click. When they
would be comfortable writing SPSS syntaxes, I'd introduce R to show how
easily and elegantly they could achieve the same goal-oriented tasks in an R
environment. However, they could still use some powerful features of SPSS in
terms of OLAP cube processing, and data preprocessing and interface R with SPSS.
My two cents,
Arin
>Message: 1
>Date: Thu, 12 Aug 2004 12:19:30 +0200
> From: "Rau, Roland" <Rau@demogr.mpg.de>
> >
> Yes, I do know the R-Commander. But I did not want to give them a
>GUI but rather expose them to the command line after I demonstrated that the
>steep learning curve in the beginning is worth the effort for the final
>results.
>
> That is why I wanted to ask the list if anyone has faced the same
>situation to persuade students to use R. Are social science students most
>impressionable with some nice graphs (e.g. filled.contour) or will they get
>a more positive attitude if I used the "R as an overgrown
calculator" like
>in Peter Dalgaards book? Or should I write an SPSS script to perform a
>certain task and demonstrate how easy, compact, and elegant it is to fulfill
>the same job in R? Just telling them "We will use R during our
course"
>without any explanation would be not a good choice in my opinion.
>
> As I have written before: I would like the students to trust me that
>it is worth to invest some extra energy in the beginning. I do not expect to
>receive any prepared demonstration from anyone of you. I am more curious
>about your teaching experiences and how you got people enthusiastic to use
>this software.
>
> Thanks,
> Roland
[[alternative HTML version deleted]]