3 ms·
Keep in mind that correlation doesn't prove causation, but it does go along with it. Any time you find correlation, something goes along with it. Maybe A caus
by saalweachter 6y ago
Keep in mind that correlation doesn't prove causation, but it does go along with it. Any time you find correlation, something goes along with it. Maybe A causes B. Maybe B causes A. Maybe A and B have a common cause C. Each of those is interesting, even if you don't know which is which from just the correlation.
Consider if the data had fallen out the other way: if there was no correlation between these measures of harm and cannabis use, that would have been evidence of a lack of harm from cannabis use.
So even if you can't draw any strong conclusions from this study, it's still an interesting data point.
- pintxo 6y agoIt might also just be a coincidence like the correlation of the "Per capita consumption of cheese (US)" with the "Number of people who died by becoming tangled in their bedsheets" [1] [1] http://tylervigen.com/view_correlation?id=7 http://tylervigen.com/view_correlation?id=7
- saalweachter 6y agoEh, correlating two time-series trends is almost a gimmick. If you saw that the per-capita cheese consumption of people who died by becoming tangled in their bedsheets was twice or ten times that of the baseline population, that could still be a coincidence, because hey, that's always a possibility, but it would be a much more suggestive correlation than just observing that any two monotonic time series correlate.