Hacker News new | past | comments | ask | show | jobs | submit login

To the extent that other studies have found such links in the past, I'm curious how much is due to pure correlation: ie, people who own cell phones and also engage in other behavior that increases the risk of cancer. This would also be much more relevant in the days when only a few early adopters owned mobiles, as opposed to now, when practically everybody does.



Valid point, but this concern should be somewhat alleviated by the massiveness of the study. I imagine that in 360,000 people, "pure correlation" should be relatively uniformly-spaced, meaning that there's enough people for all "correlative" behaviors to be statistic noise.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: