Mann-Whitney U Test: Theory and Tutorial in Excel

[@kaylaquijas123]

I like how it's titled "Mann-Whitney U Test", but the video is demonstrating a Wilcox matched-pairs test. Helpful.

[@charliebaker5731]

Are the group's dependent?? I didn't think so but idk lmao

[@charliewoods219]

literally!

[@superspacemonkey1684]

Thank you so much for this tutorial, I'm at university first year studying Psychology and I've been pulling my hair out trying to consolidate raw data from my first experiment. You really made it so much easier to grasp, keep up the hard work :)

[@nolapickering1917]

Fantastic video, thank you. Using this for a huge HS research paper, and I attempted to use SPSS, but there are no good beginner tutorials I could find for the procedures of this particular test! This was far less frustrating and far more fruitful. You're a king

[@darahox]

Thanks for this! Very helpful! You comment in the video you are not sure how to number sequentially easily- type a 1, enter, type a 2, then select both 1 and 2, then grab corner dot and drag down, it auto-populates the sequence as far down as you want

[@paulfitzgerald4625]

So at 8:20 even though you have several values for 12, 14 and 15 you still rank them in ascending order? Why don't the two 12 values share rank 8 for example?

[@sambitsarkar7619]

Sir i have a confusion , i guess that if ustat< ucrit, then we should accept the null hypothesis.But in your example you told that if ustat< ucrit , then we reject the null hypothesis. Please clear me if i am wrong.

[@arnolddistor5445]

Yes, I agree with the things that you have mentioned here, remember that if a U Statistics is less than the Critical Value then, accept the Null Hypothesis since the U Stat falls within the Acceptance Region. Thus, there is no significant difference between those two groups. Thanks.

[@JyotiSharma-yx5ux]

Thank you so much for this video explaining the concept so clearly. I have been able to use Mann Whitney U Test and Kruskal Wallis H Test. Now for my thesis, I want to cite some published source for these tests. Can you please share link for these tests where concepts and interpretation of results of these tests is given? Thank you in advance.

[@shaz_kafshaz]

I'm having to do this for my Environmental Studies assignment and just a little tip; You can do `=SUMIF($D$2:$D$100,1,$F$2:$F$100)` and `=SUMIF($D$2:$D$100,2,$F$2:$F$100)` in case you ever changed the sorting of those tables. That way it'll still sum up all of Group 1 and Group 2 correctly :) thanks for the video. Let's hope I did it right for my assignment.

[@lozchapmann]

How do you find your U_critical if you have 97 pieces of data in one set?

[@sclasses1894]

u should have taken the average values of ranks for the repeated numbers. use rank.avg function of excel

[@hslade44]

Does the rank not matter if there are 2 of the same value? For example in the all data column there are 2 12s and you ranked them differently- does this not affect the outcome?

[@anthonyfranciscosr]

That's the same question I had. Why did he not average the the ties?

[@giuseppesorrentino9466]

@@anthonyfranciscosr he should have done this.

[@k-zone2753]

yes, I have the same question, but mabbye it is because it takes too long with long datasets so he choose to skip it. but it will have an influence on if you can reject the hypotesis or not.

[@AutOdometer]

Thank you. I am completing my LSSBB and want to know how to do all the parametric and non parametric in excel.

[@ethanmoodley4080]

literally saved one of my most important High school papers!!!

[@CatalystUniversity]

Glad this was here to help!

[@Zikoo08938]

@@CatalystUniversity after this is it possible to get a p value?

[@yannilublin6455]

Are you sure about the final conclusion? If the test value is less then critical value, then we fail to reject H0. It means there is no significant difference between means.

[@keysky_1622]

Lucky that I can reply to your question relatively early. As I have found out, Mann-Whitney U test proves to be an exception. Check kzbin.info/www/bejne/eIWUd36aZratoNk, here the case is the same.

[@keysky_1622]

to your convenience, it is 8:22 in the link

[@sambitsarkar7619]

@@keysky_1622 Thank you so much brother. Just one question ... is this exception be the same for wilcoxon signed rank test as well?

[@keysky_1622]

Great question! Check this out: sphweb.bumc.bu.edu/otlt/mph-modules/bs/bs704_nonparametric/BS704_Nonparametric6.html I think the cases are similar for rank-tests. The calculated test value should be less than the critical value for us to reject the null.

[@Sara-kz3bs]

How do I find the p value based off the U statistic?

[@mursidedumlupnar3555]

What if my N2 is more than 20?

[@coconut5486]

yeah same my N2 is 21 and I can't find any critical value table higher than 20!

[@once4all125]

What happens if the number of samples are less? Or if my data is not enough?

[@matinramezani9927]

thank u sir for ur exelent explanaition

[@amanbundhun1265]

Hi, What if the sample size is 200?

[@avachavoshi574]

really useful, thank you :)

[@chelseygreene3571]

Thank you!!!

[@SisilArd]

Thank u so much kakak..

[@mirekjurasek6093]

/