I find this topic hard to comment on specifically, but I think you are probably on the right lines about specificity of the observed phenomenon being of key her

I find this topic hard to comment on specifically, but I think you are probably on the right lines about specificity of the observed phenomenon being of key here. Using a ridiculous example "for a sample set of 10, the author found that exposure to fire causes burns" would hopefully not draw similar sampling concerns but as I don't know this subject matter deeply, I can't comment whether it is true for this case