EzekielRaiden
Follower of the Way
Ooh, so it's super vulnerable to things like Simpson's "paradox". That is, the situation where treatment A does better with low-severity patients, and treatment A does better with high-severity patients, but when combined all together, treatment B does better overall. Not saying that this definitely 100% guaranteed happened--but this kind of thing is a perennial problem when you aggregate data that obfuscates the underlying source data.It’s a survey across a sample of North American hobby stores. I don’t know what algorithm they apply to the responses. But it’s not sales figures. It’s possible that the #3 thing on one store’s list sells more than the #1 thing on another but that level of detail isn’t provided.