[comp.edu] Value of a Course in Statistics

duncan@dduck.ctt.bellcore.com (Scott Duncan) (03/16/90)

In article <211@samna.UUCP> jeff@samna.UUCP () writes:
>
>             Get rid of statistics.  My own experience (and I've talked
>to a few others who agree) was that statistics consisted of learning
>a whole bunch of silly formulae and how to plug your numbers into them.
>There didn't seem to be any (hmm ... rather than "any", how about
>"enough to make it worthwhile") generally useful principles to be learned.
>
>I considered this to be a complete waste of time.  I don't remember
>any of the formulae now - in fact I didn't remember them three months
>after the course was over.  So, if I ever need to use any of the formulas,
>I'll have to go look them up anyway and I could have done that just as
>easily without wasting a semester on the course.

Having recently taken a self-paced course in statistics -- videos, texbook,
and workbook -- I can appreciate that this might be the case after a course
in statistics.  My impression at the end of the 30 hours I spent taking the
course was that one needed a computer-based statistical package handy to
avoid the drudgery of manual computation and emphasis on formula memoriza-
tion.

I thought the course I took was a decent intro to the subject, however.  The
videos were based on a PBS-like series of videos (26 30-minute tapes) which
emphasized how statistics were actually used.  Yet, the variety of forumlae
presentated got rather confusing after a while.

I was wondering what other people's experiences were with taking statistics
courses?  I took this one in order to get a better apprecitation for the
kinds of methods and analysis done in software engineering studies, i.e., so
I could begin to understand what people's results might mean, including the
shortcomings of certain computations and techniques.  (I don't think this
course that I took completely did that, but it's aimed me, I think, in the
right direction.)