The analyst, looking at data from a process, faces two problems:
- Searching for trouble that doesn’t exist
- Overlooking trouble that does exist
The statistical methods used by Western Horse Watchers are designed to minimize the losses from these mistakes. There are no probabilities, no underlying distributions.
The formulas help you decide, by calculation, if the results from a process should be attributed to chance or whether they should be associated with one or more assignable causes (nameable, knowable, findable).
Do these data look like they were produced by a simple random process?
If you’re talking about the Assateague herd, the answer is ‘No.’
Data from the recently concluded gather at Nevada WHR look like they were.
This doesn’t mean that wranglers weren’t on horseback cutting mares from the herd, to skew the sex ratio, there’s just no evidence of that in the data.
RELATED: A Lesson from Industrial Quality Control.