At 10:16 AM -0500 12/22/00, Barbara Watters wrote:
>I firmly believe that asking students to show their steps in the
>computation of descriptive stats, correlations, and t-tests is _not_
>too much to ask. Later on, when we begin the more difficult
>statistics, I find that students are much more likely to be able to
>comprehend them, and more importantly, communicate verbally and
>effectively about their meanings. I teach students how to use SPSS
>later in the course; reading and interpreting the printouts is a
>challenge for many students.
>
>I am very interested in other TIPSters views on this. "How dare I
>ask students to show their work on math problems!" Indeed...
>
>Barbara Watters
>Penn State Erie, The Behrend College
My preference is to show student how to compute them in excel, using
the definitional formulas (NOT using the excel functions). They
learn excel (a useful tool) and they learn how the formulas work. We
can then concentrate on why they work (e.g. why the deviations around
the mean add to zero, why larger samples give smaller standard
errors, why t-tests are ratios, etc.).
I am not sure that the act of adding by hand produces any effect of
itself. It is setting up the problem and paying attention to the
details. Students must do this in excel, but do not have to do
simple computations.
-Chuck
- Chuck Huff; 507.646.3169; http://www.stolaf.edu/people/huff/
- Psychology Department, St.Olaf College, Northfield, MN 55057
