Re: reality vs. science fair vs. hypothesis testing

[John Denker <jsd@MONMOUTH.COM>]

At 11:05 AM 9/24/99 -0400, Michael Edmiston wrote:
> my daughter ... wanted to learn how amplifiers work and
> then try to build an amplifier for her guitar.  That type of project is
> absolutely not allowed in our local science fair.
...
> If the project does not have a hypotheses, test, accept/reject
> hypothesis format, the project fails.
...
> It's their whole grade for one grading period.

This scenario mixes up several wildly divergent concepts:
        * How is real science done?
        * What makes a good science-fair project?
        * What is the role of hypothesis testing?

When Galileo built his telescope, he was not testing the hypothesis that
there were moons around Jupiter.  When Kamerling Onnes decided to measure
some real cold metal, he wasn't testing a hypothesis about
superconductivity.  Mendeleev spend many years playing with funny-colored
stinky precipitates before he even dreamed of formulating a hypothesis
about periodic tables.

I object to the term "the scientific method" because it implies that there
is only one method that scientists use.  In fact we have oodles of
different methods.  Hypothesis testing is one of them.  It tends to be
useful relatively late in the game, when we mostly know what is going on.

Early in the game you have to just try stuff.  In so doing, you run the
risk that you might not learn anything worth reporting -- you might be
looking under the wrong lamppost.  In the lab where I work, only about 10%
of what we do turns out to be useful.  The trick is to manage it so that
the successes are big enough to pay for everything including the failures.

Presumably the high-school teacher doesn't want to flunk 90% of the
class... which automatically implies that picking science-fair projects
will be *totally* unlike picking real research projects.

There's more to life than school.  If the kid wants to build amplifiers,
she should build amplifiers.  Maybe after building three or four amplifiers
and reading three or four books on the subject, she will *then* be in a
position to make interesting hypotheses and test them.

The remaining question is, what should she do for the science fair?  Given
the foolish policy of linking it to the grade, the project must be
non-risky.  It will not be research, but it still has to look like science.
 She should pick something she basically already knows how to do.  Or
perhaps something her illustrious father knows how to do and can help her with.

For example, here's a low-risk amplifier-related experiment that should be
within the capabilities of a 13-year old:  Predict how long a speaker wire
can be before it materially degrades the sound.  Predict how this depends
on wire gauge.  Do the experiment.  (The devil is in the details.)  Comment
on whether fancy large-gauge cables are really worthwhile or just a load of
hype.

Reading the list of projects from previous years might give you other ideas.

______________________________________________________________
John S. Denker                                jsd@monmouth.com