6:40 why is that a large sample size will differ from normality? In learning statistics, I thought that large sample size is always better. 2nd question: Does it mean that I can no longer perform Linear Regression and ANOVA using these variables that are non-normal?
@ResearchByDesign
4 жыл бұрын
A large sample size is the BEST WAY to get a representative sample, but randomness means that it is not guaranteed. Yes, bigger is better, but not a guarantee. In linear regression, it is important that your DV be normally distributed, but there is no normality assumption for the IVs. For ANOVA, it is desirable to have a normally distributed DV, but you can still use and learn from a non-normal DV. The assumption of normality is really about the distribution of sample means, not a particular sample
@GhennadiiGubceac
6 жыл бұрын
Great tutorial! Thank you very much!
@taratehrani8019
6 жыл бұрын
Tnq soooo muchhhh
@elvinthomas4172
4 жыл бұрын
should we remove the outliers ?
@ResearchByDesign
4 жыл бұрын
You can remove the outliers or you can scale them back (Winsorize) them to the highest reasonable value.
Пікірлер: 8