[CLUE-Talk] Introduction to Computer Science Course
Jeffery C. Cann
jccann at home.com
Tue Feb 20 18:19:15 MST 2001
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Tuesday 20 February 2001 11:06, Wayde wrote:
> When it comes to "testing" there is little training with respect to
> experimental design. In general it seems that the common wisdom taught is
> to control everything and only change one variable at a time. While this
> isn't grossly an incorrect procedure it makes the assumption that the test
> variables are independent from one another and not correlated. This is a
> very special case in general. More often than not, the effect of one or
> more variables will depend on the values of the others. The net upshot of
> this is that the oft touted one-variable-at-a-time testing method will
> fail on real world problems. The solution may in fact be to change more
> than one variable at a time.
I agree. When you get into advanced training (in my M.S., anyway), we
learned how to do multivariate experimental design and results analysis.
- From a teaching perpective, the first step is to change one variable so the
concepts of testing can sink in.
One of the biggest things I used to tell my Bio 101 students is to beware of
correlation studies. For example, we know that yellow teeth results from
smoking and we know that cancer also results from smoking. Does this mean
that yellow teeth cause cancer? No, but reading AP stories about the latest
gene that causes { homosexuality or alcoholism or criminal behavior }, the
dangers of correlation studies (single variate) become problematic.
My point is that all students (not just intro to CS students) should learn
how to think critically before accepting scientific facts because facts are
usually hypotheses and sometimes tested theories. However, new studies shed
new light on facts; thus, a 'scientific fact' is really a fluid concept...
Jeff
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: For info see http://www.gnupg.org
iEYEARECAAYFAjqTF5cACgkQw3/GBQk72kBY3wCdHH1KZfeE2kCmXNhc9pyDZpov
3JoAn3EsN4HJvRf8l/Um0kiXNTEhpI3M
=Xjcm
-----END PGP SIGNATURE-----
More information about the clue-talk
mailing list