[QUOTE=MelantheLLC;6315653]
What other reasons is it not a good study, 3dogs?
Silvia and 3dogs, are you saying the study has no value, or that the conclusions in the abstract are overdrawn?
I thought that anything less than 30 was considered a “small sample” rather than completely invalid. Performance comparison p-values are pretty low, based on blinded judging. Not sure what the control would be in this case…not a third training program, so just horses not being trained?
How would you design this study more correctly?
ETA: I’m not a supporter of Monty Roberts, in fact never seen him work, though I’m watching the vid that’s on eurodressage now, if I could ever get it to stream more than 5 secs at a time. However, I’m intensely interested in how horses learn.[/QUOTE]
In terms of doing studies, less than 30 is a small sample, and less than 10 doesn’t offer a statistically significant result.
Some suggestions I might have considered to make it more significant:
10 horses for trainer 1
10 horses for trainer 2
10 horses for control
Two methods of assessment were used, heart rate and performance review.
Heart rate is objective, this was the most relevant data of the test (but no control to compare it to).
Performance review is subjective hence needing judges. A more appropriate test might have been time taken to negotiate a course (vs heart rate)
The tests could have been done by the same person (ie a 3rd control person) to eliminate differences in riding as opposed to the actual training methods.
20 days at 30 minutes per day is also a very short period of time to train any horse to go under saddle and negotiate a course/test.
Controls used on the group could have included testing the horse’s resting heart rate and under stress to compare against results during the tests.
Just a few thoughts? I am not an expert on these things!