Date: Fri, 5 May 2006 18:23:22 -0400
Reply-To: Arthur Tabachneck <art297@NETSCAPE.NET>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Arthur Tabachneck <art297@NETSCAPE.NET>
Subject: Re: Speaking of training wheels
First, definitely follow David's advice. But, as an ex-institutional
researcher, I'd also add to ensure that you focus in on the various tasks
that you will repeatedly need to accomplish.
For example, institutional researchers continually have to read in data
from numerous sources (e.g., data step, proc import (all types),
sas/connect, libname engines, etc.), combine and/or merge various data
sets (e.g., using merge within a data step, proc sql, update, etc.),
summarize data and pass the results into whatever you happen to be doing
(e.g., proc means, proc summary, using macro variables, using SCL), doing
all kinds of analyses (e.g., proc whatever statistics you think you will
need), doing all kinds of summary reports (e.g., proc report, tabulate,
means, freq), reviwing data (e.g., insight, descriptive, and many of the
already mentioned procs), and knowing which functions might help you in
any of the above (i.e., RTFM).
Of course, give yourself plenty of time. I've been trying to accomplish
all of the above for the past thirty some years.
On Thu, 4 May 2006 11:49:45 -0400, sean maru <maru859@GMAIL.COM> wrote:
>Now that I have completed my little project I would like to continue my
>study. My thought was to take the same data set and rewrite my program a
>times and experiment with different types of analysis and different
>approaches to the coding (hopefully to make it more logical and more
>efficient). Mainly I know i need practice but I don't want to pursue some
>irrelevent tangent. If you were to devise a "basic training" course for a
>complete newbie what would it include? How would it be the same or
>from the way you learned, etc?
>The type of work i expect to do will be in the vein of institutional
>research. Therefore there will be lots of issues regarding reconciling
>from several different databases (different types) and comparing every
>metric in the world as a function of demographic variables (race, gender,
>etc.). In this kind of work i imagine the statistical demands are not high
>but being able to code efficiently will be a key I think.
>Any thoughts would be appreciated.