After taking in an overview of my yard from my second floor window, I’ve declared it a scientific fact my dog takes a shit in whatever damn direction she feels like.
The paper is maybe a bit more scientific than that, but I’d really need to go through the statistics carefully before believing the results.
It’s a very suspicious sign when the results are “Not significant overall, but this particular subset shows a significant result!” You take 20 subsets and, chances are (if you use inappropriate statistics) you’ll find one that has a significant result even if all the numbers are completely random.
I noted this years ago back when we had 2 black labs. Actually noticed several things.
When one pooped, the other did also at the same time, or right after.
The both did their duty facing the same way.
They always faced the same general, but not exact, direction. Never gave the direction a thought but it was indeed generally north.
I just heard a discussion of the paper by a number of scientists. Their take was that a massive amount of data was collected. When the primary data showed nothing, they began comparing other things. Massaging the data, you often will coincidentally find patterns if you look long enough.
My Catahoula craps facing uphill, for pretty obvious reasons. Pees that way, too—she hates it when the pee rolls back onto her toes.
Isn’t that the same thing as confirmation bias?
We Have Confirmation!
Reproducibility Has Been Achieved!
OK, we have 3 N, 2 E, 2 W, and 1 S. I’m not seeing any trend, here, yet. Any PhD type dopers want to weigh in as to how many shits I have to record before we have good data?
The dog will keep shitting, but I’m wondering how many I have to record before we have decent data…