That would be because the properly run surveys with good protocols are all paywalled by academented journals.