Is it possible to have a very significant Chi-squared test, but a terrible correlation (I'm talking around 0.2)??
Asked
Active
Viewed 30 times
0
-
This question has come up many times across different tests vs measures of effect; the answers are always pretty much the same ... e.g. https://stats.stackexchange.com/questions/114428/how-do-you-interpret-a-significant-but-weak-correlation (which is also related to the chi-square and seems to be close enough to call a duplicate) or https://stats.stackexchange.com/questions/104082/huge-sample-sizes-and-null-hypothesis-testing or https://stats.stackexchange.com/questions/273215/trivial-effect-size-but-statistically-significant ... ctd – Glen_b Jul 07 '17 at 03:28
-
ctd ... (in short "yes, if the sample size is large enough", and "statistically significant is not the same thing as practically meaningful") – Glen_b Jul 07 '17 at 03:33
-
Also see https://stats.stackexchange.com/questions/245039/r-linear-regression-very-small-coefficient-and-r-squared-but-significant-p-val and https://stats.stackexchange.com/questions/226477/relationship-between-effect-size-and-statistical-significance and https://stats.stackexchange.com/questions/229828/what-is-the-difference-between-clinical-significance-and-statistical-significanc – Glen_b Jul 07 '17 at 03:37