Date: Mon, 7 Oct 2002 11:29:11 -0400
Reply-To: Peter Flom <flom@NDRI.ORG>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Peter Flom <flom@NDRI.ORG>
Subject: Re: Can SAS be dangerous to your job?
Content-Type: text/plain; charset=US-ASCII
Indeed, I think we should not blame the tool at all; indeed, giving that
one works at a place that does not want to be informed when they make an
error, or of ways of fixing those errrors more quickly, getting laid off
might be good for one's career, at least in the long term.
But the analogy of SAS being like a loaded gun in the hands of a child
isn't true only of curious programmers. It's also true of inadequately
trained people doing data analysis.
e.g., I was a TA in graduate school (in the Psych department). I got a
reputation for knowing statistics. One colleague asked me why her
output was so strange, in that an analysis she was running gave a
p-value of 1.
After repeated questioning, I finally got her to show me what she had
done.
It was, roughly
proc glm;
class IV1 IV2;
model dependent = IV1 IV2;
by IV1;
run;
I pointed out that she had done an ANOVA on a variable she had
controlled for. This produced a blank look. Then I told her that it
was an analysis of variance on a variable she had controlled for.....the
light didn't dawn. So, I said
"Youre comparing boys and girls, but you're doing it seperately for
boys and girls!"
Peter
>>> Ian Whitlock <WHITLOI1@WESTAT.COM> 10/07/02 11:01AM >>>
Shortly before NESUG I received a message containing:
>>>>>>>>>
I used to try to write SAS code (I had the only copy in the company) to
try
to cleanup and identify problems but upper management decided I was not
a
team player and layed me off.
<<<<<<<<<
I showed the message to someone else who quickly replied pointing to
the
above sentence.
>>>>>>>>>
I did exactly that -- used SAS to identify problems when we were
building
the XXXX (data entry and editing system) for YYYY. I also got laid
off.
hmmmm....
<<<<<<<<<
The juxtapositon of the two statements got me to wondering about
others.
Perhaps one should not blame the tool. On the other hand, placing SAS
in
the hands of a curious programmer is a little like giving a child a
loaded
gun.
Any comments or anecdotes?
IanWhitlock@westat.com
---------------------------------------------------------------------
The contents of this message is the work of the
author and does not necessarily represent the
opinions, recommendations, or practices of
Westat.
---------------------------------------------------------------------
