The semi-inevitable downvotes may be because xkcd links are seen as cheap canned injects that although relevant and often funny don't really add anything novel to a debate. And perhaps - the horror! - they might not actually always be correct in a specific case.
Probably because someone posts "correlation does not imply causation" as a rote "I am very smart" reply to any published study, regardless of the actual contents and claims of the study.
