Timeline for What are the worst (commonly adopted) ideas/principles in statistics?
Current License: CC BY-SA 4.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 22, 2023 at 1:45 | comment | added | Glen_b | This doesn't imply what you said is wrong, but I fear people might be led to misunderstand what you were suggesting. Readers beware: just because you find a big correlation (like 0.98, say, even in a very long series) doesn't mean there's causation at all. | |
| Dec 22, 2023 at 1:42 | comment | added | Glen_b | @DavidErnst I'm including this caveat to warn people not to misinterpret/over-interpret what you're saying: Take two binary random walks ... (e.g. take two distinct coins, and toss each repeatedly and let the values of each series be number of heads minus number of tails seen so far on the coin generating that series). As the sample size (number of tosses) becomes large, typically the size of the absolute correlation between the two series will tend to be quite large. [This is an example of spurious correlation among nonstationary series that are not in any way causally connected.] | |
| Jul 27, 2020 at 11:47 | comment | added | David Ernst | Correlation precludes causation is an even worse misinterpretation. But even saying that true correlation (not a type I error) can happen without some kind of direct, indirect, common cause or inverse causation is wrong. You don't know much about this causation, but you do know that it is there somewhere | |
| S Jul 24, 2020 at 12:27 | history | answered | Dave | CC BY-SA 4.0 | |
| S Jul 24, 2020 at 12:27 | history | made wiki | Post Made Community Wiki by Dave |