Non parametric tests work great with small sample size data. Now suppose I have a large sample size of 10 million samples that are not normally distributed. Will non parametric tests work in this case?
$\begingroup$
$\endgroup$
5
-
1$\begingroup$ All commonly-used tests work better with more data, no worries there. $\endgroup$– jbowmanCommented Jan 29, 2019 at 19:01
-
$\begingroup$ What do you mean by "work great with small samples"? $\endgroup$– Joel W.Commented Jan 29, 2019 at 20:35
-
1$\begingroup$ Could you please be a bit more specific about why you think that non-parametric tests might perform poorly with large data sets? $\endgroup$– EdMCommented Jan 29, 2019 at 22:20
-
$\begingroup$ I was doubtful that non parametric tests only work for low sample size. $\endgroup$– Rishi SharmaCommented Jan 31, 2019 at 18:56
-
1$\begingroup$ I'd add a rider: Most tests are irrelevant with very large samples. If work well means "work as designed if conditions are appropriate", the answer is Yes. If it means tell you something useful, the answer is often closer to No. $\endgroup$– Nick CoxCommented Jan 31, 2019 at 19:03
Add a comment
|