Date: Thu, 9 Dec 2004 10:03:19 0500
ReplyTo: Paige Miller <paige.miller@KODAK.COM>
Sender: "SAS(r) Discussion" <SASL@LISTSERV.UGA.EDU>
From: Paige Miller <paige.miller@KODAK.COM>
Organization: Eastman Kodak Company
Subject: Re: SAS IML and SAS GRAPH  ease of learning? comparison with R?
ContentType: text/plain; charset=usascii; format=flowed
Peter Flom wrote:
> Hello
>
> I am currently involved in a new project for me, a metaanalysis.
> Currently, I use SAS BASE (a little), and SAS STAT (a lot), for fancier
> graphs and/or statistics that aren't in SAS STAT, I use R.
>
> This project will necessarily involve me learning more SAS BASE. But
> after that we have a choice, and I'd like some advice
>
> We can either 1) Get SAS GRAPH and SAS IML or 2) Stay with what we are
> doing. SAS GRAPH and IML are expensive; R is free. But R has a steep
> learning curve, and some of the analyses may need to be done by other
> people than me; those people don't know R, and don't want to learn it.
> With SAS, we can rely (to an extent) on programs that Wang and Bushman
> wrote about in their book on metaanalysis with SAS.....But at least one
> of us (me) will probably have to learn at least some of how those
> programs work, in case they need to be modified (which seems likely).
>
> Sorry for going on so long, but (as people who've been following other
> threads know) I like to provide context.
>
> Any advice will be appreciated.
I have never learned R. I did at one time look at it and I decided
that since I already knew SAS/GRAPH and SAS/IML, I felt little would
be gained from learning R.
You are in a different situation, where you have not learned
SAS/GRAPH and SAS/IML; but I think the learning curve is much less
steep than the learning curve to R will be, because you already know
SAS/BASE and SAS/STAT. Learning most of SAS/GRAPH is no different
than learning a new procedure in SAS/STAT; there are a few new
concepts in SAS/GRAPH that don't have an analogous concept in
SAS/STAT such as AXIS and SYMBOL statements, but these I don't think
these are hard concepts. Learning IML is similar to learning new
features of the data step, where now you can perform matrix
manipulations on your data; again, if you can program the SAS data
step and you understand matrix manipulations, I don't think IML is
particularly hard to learn (if you have never been exposed to matrix
manipulations than IML and any other matrix language will be
inscrutable).
It is also quite nice to have one program, in this case a SAS
program, that does all the analysis. If you want to do a complicated
analysis in IML, then output the results to SAS/BASE, SAS/STAT and
SAS/GRAPH for further analysis or summarization of results, you can
do this in all one program. No need to write a SAS program and an R
program and link them together somehow.
Of course, only you and your company can make the tradeoff of these
advantages against the cost of adding modules to SAS.

Paige Miller
Eastman Kodak Company
paige dot miller at kodak dot com
http://www.kodak.com
"It's nothing until I call it!"  Bill Klem, NL Umpire
"When you get the choice to sit it out or dance, I hope you dance"
 Lee Ann Womack
