Re: [RC] rider weight studies, part 2 - Sisu West Ranch
"...The last time I was at Tevis collecting data for a
different study, I also collected rider weight, horse weight, condition scores
and a couple other parameters, and just for fun, predicted who would finish and
who would not, based solely on the predictive values derived from the above
data. ..."
An 88% or 94% prediction from
a multiple correlation of the above parameters makes sense to me. All of
these have to be important.
Could you point me towards
the reference for the coeficients deviations etc.? I might still remember
enough to convince myself about the validity of the conclusion we have been
discussing.
Everyone please understand, I
was born in NJ. As a result I am skeptical by nature. There are two
common errors talked about in Statistics classes. (I can never remember
which is type 1 and which is type 2) I am much more likely to reject a
hypothesis that ultimately turns out to be true than to accept one that
ultimately turns out to be false. Luckily in my working life, it was
almost always possible to do one more controled experiment and thus sort things
out before a freight car load of bad product was made.
Ed
Ed & Wendy Hauser 2994 Mittower
Road Victor, MT 59875
