Stephen wrote:
> Rick Adams disputed my top four negative correlations with shoe size
> (steps wall-to-wall in the classroom, distance from the head to the
> ceiling, length of hair, and number of teeth) by telling us that he's
> a tall, hairy, dentured person with tiny feet (or something like
> that).
Nope.
No dentures! :)
> Ah, Rick, Rick. Aside from the fact that you've given us _way_ more
> information about yourself than we're comfortable with, you forget
> that we're talking correlation here, and given enough subjects, even
> a teensy r can be significant. And we can always throw out outliers,
> anyway (not that I'd dream of calling you an outlier, of course. This
> is never advisable with anyone who's 6'4").
But the whole point is, we're _not_ talking a large subject base here.
You've lost sight of the original question which concerned a demonstration
that could be conducted _in a classroom._
In that case, the subject base is very limited and a single outlier can
destroy the entire demonstration for the instructor--and as you well know,
when trying to perform such a demonstration in a class there _always_
seems to be an outlier present (look at this group alone, I was able to
refute two of the four [head to ceiling & length of hair] personally, and
I have no doubt there is at least one person on the list with dentures and
another with arthritis or some other joint disorder that would necessitate
small steps). That outlier will weaken the demonstration for students who,
unlike ourselves, are not as clear about the concepts of correlation and
the fact that an outlier doesn't invalidate the results.
> (I actually think the one about hair length and shoe size shows
> promise. I hope someone tries it. Just make sure you have a large n
> and lots of short women in class.)
And no long haired men, of course. Or tall, long haired women who may be
paired with "vertically challenged" short haired women. :)
Rick
