Search All of the Math Forum:

Views expressed in these public forums are not endorsed by NCTM or The Math Forum.

Notice: We are no longer accepting new posts, but the forums will continue to be readable.

Topic: "standard computer output"
Replies: 0

 Peter Bruce Posts: 10 Registered: 12/6/04
"standard computer output"
Posted: Sep 13, 1996 10:53 AM

The use on the AP exam of "standard computer output" (whether minitab or
other) raises a problem for those contemplating a resampling/simulation
approach to statistics.

Resampling procedures (Monte Carlo, permutation, bootstrap) can be used to
solve the same problems that formulas (t tests, chi-square tests, etc.)
solve, except proceeding from a more general base (specify the population of
interest, constitute a simulated representation of it, draw "resamples",
observe the behavior of those resamples).

This gives students an easily understood tool for solving problems (not just
illustrating them). Our work has shown that students do better, understand
more, retain more and have a more positive attitude towards statistics when
taught resampling than when taught conventionally. And there is no
theoretical reason why resampling (in and of itself) should not serve as an
appropriate alternative (or supplement) to t-tests, etc.

The AP exam's requirement that students interpret "standard computer
output", however, is likely to discourage AP teachers from even considering
a resampling/simulation approach (whatever its technical and pedagogical
merits) -- resampling is seen as an added burden that must be borne.

If it were universally agreed that "standard computer output" of canned
tests (t-tests, etc.) were an essential feature of understanding statistics
AND WOULD REMAIN SO, this would not be an issue.

This is not so, however -- resampling can get the job done (and I would say
better) without those formulas.

If I'm right, how will we ever know unless students are allowed to learn
this way? And what teacher is going to teach them this way if they must
learn "standard computer output" anyhow?

I recognize the desire to create an exam that tests what are commonly
regarded as the "must know" features of a field. The question is "When
innovation occurs (in this case, the advent of computer-intensive resampling
& simulation methods), how do you keep the exam itself from becoming an
obstacle to integrating that innovation into the classroom?

Peter Bruce

Disclosure: I have taught resampling statistics at the University of
Maryland, and am a principal in Resampling Stats, a small firm that makes
software to do resampling.

>
>I have a question about the use of Minitab output on the AP Statistics
>exam.
>
>Now that the TI-83 is available, it is possible to teach a
>"technologically correct" introductory statistics course without
>using a computer at all.
>
>How important it is for students to use Minitab or see Minitab output,
>since such output may appear on the exam? For most statistical tests, the
>TI-83 output is different in appearance from Minitab, although the important
>values are usually provided in both settings.
>
>Would it be reasonable for an AP Statistics course to be taught using the
>TI-83 exclusively, or would the students pay a penalty on the exam for
>not having been exposed to Minitab?
>
>How would the answer to these questions change if the word "Minitab" is
>replaced by "some statistical software package" or the name of a specific
>statistical software package other than Minitab?
>
>
>Tom Short
>Villanova University
>
>--
>
>+--------------------------------------------------------------------+
>| Thomas H. Short |
>| Department of Mathematical Sciences Phone : (610) 519-6961 |
>| Villanova University Fax : (610) 519-6928 |
>| 800 Lancaster Ave. E-mail: short@monet.vill.edu |
>| Villanova, PA 19085-1699 |
>| |
>| World Wide Web: http://renoir.vill.edu/~short/ |
>+--------------------------------------------------------------------+
>
>