|
I
doubt that we could actually garner any significant statistics from the pull
codes. (Truman - you're the mathemetician... what do you think?) The biggest
enemy of statistical signficance is variation, and just consider the amount of
variation there is with each L pull: The obvious ones are: state, region,
temperature, terrain, speed, horse, rider, vet, type of lameness, cause of
lameness, - then you should really consider - time of year, age of horse,
age of rider, fitness of rider, pace, feed, shoeing method, conditioning level,
rider experience, moon phase... etc. Having done research and tried to glean
statistical significance from my work, I know how difficult it is. It would be
very hard to (for example) find a statistically significant correlation between
things that seem rather obvious - speed/temperature/Metabolic - or
speed/terrain/Lame. If we were to eliminate all of the other variation factors,
we might be able to say ' look, when you ride too fast on rocky terrain, your
horse has a greater likelihood of becoming lame'. But once we factor in all the
other contributing elements it just becomes common sense, not statistical
significance.
So
Frank... I guess I agree with you. A simple DNF on the results would be
adequate, and if AERC really wants meaningful data, we should come up with
something else, such as a horse or rider tracking system (log
book).
Steph
|
