dennis roberts wrote:
>
> At 11:35 PM 4/24/01 -0300, Robert J. MacG. Dawson wrote:
>
> > Yes. That is, in my experience, students, small dogs, and
> > most white
> > mice can use a t table, at least for values that actually turn up.
> > Interpolating loses a few students and the mice; the dogs and some
> > more
> > of the students get worried about what to do for 90 degrees of
> > freedom
> > if the table only goes to 60, or if t=16 and that df row only goes
> > to
> > t=3.2
>
> yeah ... tables suck sometime ... but, if one ACTUALLY looks at the
> table ... changing from df = 60 or df = 100 (where are the 61 to 99?)
> ... makes only a trivial 2nd or 3rd decimal place difference ...
Sorry, Dennis, guess I wasn't too clear there. My point was that
while the algorithmic use of the table seems to sink in pretty well
after one lecture & one homework assignment, facts such as those you
just mentioned are understood much more slowly and less successfully.
Similarly, certain superstitions, such as using-Z-above-30, seem to bind
to the brains of the students in the Big Intro Course, tighter than
chlorine to sodium; one mention of such a misconception in the textbook
+ half an hour of explanation as to why this is a bad idea leaves the
misconception ahead on points.
-Robert Dawson
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
