Neil Dullaghan writes:
I have a strange set of correlations and am wondering if they are due to some oddity of statistics rather than real associations, but I am quite lost as to an answer.
The study in brief: 3 independent surveys asking respondents how many [insert animal] would they trade for 1 human. e.g 120 chickens for 1 human. All surveys find that the tradeoff is consistently and strongly correlated (nearing r=1) with p
My inclination is to not take this correlation too seriously as it’s just based on a few types of animals. Add more animals and you could get a much different story. Also, I think that adding more extreme cases will increase the correlation. For example, imagine you include the Rocks category to the first graph above. Rocks have zero neurons, and I assume the tradeoff would be infinity, thus the data point for Rocks would be at (-infinity, infinity) on the log scale, and including that with the other points would pull the correlation all the way to -1.
Also, I have no idea how I would answer this survey question myself, as I don’t understand what is meant by this tradeoff. What does it mean to trade off 120 chickens for 1 human? Does that mean that if I serve chicken 120 times in a two-year period, I should kill a person to compensate? I feel like I’m missing something here?
P.S. I was told that the the survey specifically asked, “What number of chickens do you think is equal in moral value to one adult human?” etc. I’m not sure how to think about “moral value.”