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

While I agree that that particular anecdote tells us nothing -- because it's void of details -- I disagree that anecdotes categorically can't tell us anything. My take isn't anti-scientific, it is against this particular brand of scientism that believes that single examples are useless. Consider that medical scientists often use case studies to better understand phenomena, because this allows for higher resolution investigation of a single example in order to shed light on phenomena, and it addresses edge cases due to individual heterogeneity that cross-sectional data can't address well. In light of this example, I would argue the anecdote vs data is a false dichotomy if you're defining anecdote to mean "example from a single person". If an anecdote is studied properly, as it is in case studies in the medical literature, it is data. It's just a different kind of data and a different mode of scientific study to lower resolution cross-sectional studies (which are also great and necessary, but have different strengths).



>>it is against this particular brand of scientism that believes that single examples are useless

Single examples are useless for discovering generalized properties of reality.

Medical case studies are only useful in conjunction with knowledge about causative associations that were discovered through statistical analyses.




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

Search: