How To... Perform a Mann-Whitney U Test in R #95

Learn how conduct the non-parametric Mann Whitney U test to determine of two unpaired samples are different with ‪@EugeneOLoughlin‬.
The R script (95_How_To_Code.R) and data file (95_Data_File.jpg) for this video are available to download from Github at:
github.com/eol...

[@yangzhimiao9367]

Thank you so much for this video!

[@user-in2ot9mz9y]

Thanks for very fine explaination!!!!

[@TadkedaarKanpuri]

Does the warning message have any repercussions? Is there any alternative to correct for it? Thanks for the great video!

[@nurinarifah9091]

thank you soo much. Your videos help me a lot. If you could possibly explain, may I know why we have to report the smaller W-statistic?

[@JelinekFS]

I also would like to know that.

[@KnowDAOself]

super useful. thanks for sharing!

[@117chris9]

Awesome thank you

[@brazilfootball]

Great video, thank you! Can I use this test on data that ranges between 0.00 -> 0.72 (each value in the dataset is an average of 10 percentages, bounded by 0.00 and 1.00) where zero accounts for about 53% of the data?

[@prernasharma6881]

Thank you . It helps me alot 😀

[@nanna6407]

What shall I write in the script if I have more variables than 2? Another $ and the name of the variable doesn't work...

[@EugeneOLoughlin]

Hi Nanna,
The Mann Whitney test is just for two variables. If you have three or more, you should use an ANOVA test (for normal data), or the Kruskal Wallis test (for non-normal data).
Dr E.

[@mylifeisinhishandsamen4167]

Can I use this for three population?

[@EugeneOLoughlin]

No - if you want to perform a non-parametric test using ranks for three (or more) populations, you should use a Kruskal-Wallis test.
Dr E.

[@zWanted579]

@@EugeneOLoughlin Hi Dr. E. would it be incorrect to perform a kruskal wallis test for 2 populations?