Re: Looking for class-testers

[Michael Bowen <fizzbowen@MINDSPRING.COM>]

At 08:03 2001/08/02, you wrote:

> I am looking for some additional class-testers for a collection of
> Web-based, automatically-graded conceptual homework items that I am
> writing.
>
> Check out the conceptual homework at www.uno.edu/~rgreene/illum.html,
> and contact me at rgreene@uno.edu for further information.
>
>         ron greene

I'm not a potential tester, but already see a small human-factors problem
that you might be able to correct.

In the section Kinematics (1D), A, 2 [Average Velocity], the applet asks
students to estimate displacement and average velocity using a rather
coarsely calibrated "ruler" and very large dots for data points. If the
ruler is calibrated only to the nearest unit, but you ask students to
estimate the data-point dots' positions to the nearest tenth of a unit,
where the dots' diameters are themselves several tenths of a unit each,
then one is likely to encounter a large number of well-meaning and
reasonable "guesstimate" answers that are off by a couple of tenths of a
unit. The quiz scoring algorithm is unaccommodating of slightly wrong
answers, even when the average-velocity answer is internally consistent
with an accompanying nearly-correct displacement answer.

In most experimental measurements, the last significant digit is "doubtful"
anyway, and shouldn't be trusted to better than plus/minus 1. What are we
teaching our students about experimental error when we insist on absolute
"precision" in reporting data, when the measuring instrument is relatively
crude AND the "spread" in each individual data point is inherently large?

The conceptual thinking behind the applets appears to have been done rather
well, in the sense that many concepts confusing to the beginning student
are addressed explicitly. I look forward to seeing the refined versions of
these "Physics Illuminations" announced in this forum at a later date.

--MB