Mann-Whitney U Test - SPSS (UPDATED)

[@kaloyanavramov1440]

Thank you so much dude. I didn't know what I was doing until I saw your video and followed the steps and suddenly I had all the answers I needed in front of me. Keep doing these videos, it seems like you know SPSS in depth.

[@YuriPavlov]

Thank you, good person! You helped me do my homework in 5 min!

[@armondo22]

Best instructional video Ive seen yet!

[@how2stats]

Thanks!

[@Fati_meow]

THANKYOUUU SOO MUCH SIR, I didnt know anything much about statisticts .your vedios helped me to complete my research project in psychology.. now i have did 30+other students research too.. tha nks to you for all vedios with great explanation...❤

[@heidilarsson1647]

Thank you, finally help out here. And a very nice voice to listen too

[@tonetty]

what if i have an experimental and control group?

[@aishwaryagautam8892]

Thank you for such a comprehensive video. I just want to know a lil more. How to interpret the z score in this test and what does that mean?

[@alejandromoraalba3543]

Guys, I don't know what i'm doing wrong but when to the "Two independent samples test" my nominal variables don't show up, only the scalar ones. Does somebody know what i might be doing wrong?

[@RahulKumar-wd1yf]

Sir , How can we show standard deviations

[@jina4030]

Dear Sir, I compared two independent groups (health science and non-health science students) by using the Mann-Whitney test, but the results showed significant with same median, is there any way to interpret this result or recommend a test that could replace it?

[@anonymous3345]

Do you have a contact, such as an email address, where I can reach out to discuss my thesis data further? I would greatly appreciate any guidance or assistance you can provide.🙏

[@MrPapakonstantinou]

Hey thanks for this informative video. I perform the Mann Whitney U test and I get my p values like ",923"... Should it be ".923"??? I guess I did something wrong

[@emadnadi152]

so the mann-Whitney U test is the same as Wilcoxon rank-sum test????????

[@joshmcnulty5950]

Yes

[@xinyee8271]

Thank you so much for your video. May I know what to do if it shows normal for male but not normal for female? Should I still run mann-whitney U test or I should go for t-test?

[@how2stats]

Mann-Whitney can handle non-normally distributed data; but it assumes the data are non-normal in the same way across groups. If I were you, I'd use the Welch's t-test (in SPSS; I have a video on that). Even better, Welch's t-test with bootstrapping!

[@xinyee8271]

@@how2stats thank you!

[@leonard2oo9]

Hi, if one group is not normally distributed and one group is normally distributed. Then which test should I do? Also, if the samples are paired do I still use the independent test for mann Whitney Test?

[@xiaotongyin2719]

Hello, I have a data which is complicated I think. So I’m confusing if I can use this test. I think the independent variables in my case are nationality, which is two countries. Dependent variables are “strategies use” . There are 20 different classifications of the strategies, and which belong to 3 macro strategies level. But my h0 is to find there is no difference between two countries in terms of strategies use across social status. There are higher , equal, lower social statutes. Analyzing utterances I collected, I defined them as appropriate strategies.

[@matteobortone6573]

Great explanation brother!

[@avivaz3038]

What is the "grouping variable" part? What numbers should be input? Should we input 1,2 regardless of how our variables are numbered in the data file? (I came across a data file in which the variables were numbered .00 and .01. In this case, how should I 'group variables'?

[@annrittukurian4913]

When am trying to do Mann Whitney group is not shown why is it so

[@annaderus1012]

You need to make them numerical, so group 1 and 2, not group A and B

[@mindworkshr]

Is it possible to use confounding variables in this test? Like if I checking for gender, can I control other variables like educational qualification and all ?

[@sharandeepbose3223]

is the r value = z value in a man whitney test? @how2stats

[@kacie.k]

What if you only have proportions for the dependent variable data? For example, 20/100 males using drugs and 5/100 females. But you don't have that all in separate excel cells?

[@dinajose7co]

Can I divide the Asymp Sig 2 tailed above mentioned, by two ? Had it been a one-tailed hypothesis? Do forgive me if I am mixing up my stats. thank you for your fab content. I'm learning slowly..

[@busraaslan2198]

You are perfect

[@magnetar2i]

Great, you've got a nice voice.

[@linearseller2835]

Since t-test is robust to non-normally distributed data, when exactly should we use MW U test? If we assume homogeneity of variance we could use t-test or MW U test, and we know non-parametric tests have lower power than parametric tests, so it seems we should always prefer t-test instead of MW U (in this context... non-normal data and homog. of variances)

[@sudiptomajumdar4210]

as you have wrongly mentioned here, you cannot run t-test for non-normally distributed data. So you are left with non-parametric tests only. you do not have an option whether it's as robust as parametric tests or not.

[@sucilestari5972]

Hw about if pvalue=0.05 sir?..?

[@how2stats]

Vast majority of reviewers won't let you get away with saying it's statistically significant.

[@saadalhumaid3959]

thank you so much your stuff are very helpful

[@viktorrenneusguthrie3542]

Great channel! Do you know if its possible to do a Van Elteren, stratified Wilcoxon , test in SPSS? If so, how? Thank you

[@taywogens3308]

HelloHello, i have data which is normal distributed, can I use the Man-Whitney-U-Test as well?

[@brensonyeo1358]

T-test is okay, unless it is heterogenous in variance, then MWU test instead

[@sudiptomajumdar4210]

@@brensonyeo1358 SPSS gives alternative t-test P values (lower set to normal values), consider that if the null hypothesis for Levene's test for equality of variance is rejected (P

[@wulidaria9288]

I think u can use t-test

[@johnbouzalas4839]

You made spss bearable

[@badrulamin9652]

what if I have 4 treatment? that's that mean that I have to perform the analysis for the first 2 treatment, then follow up for another 2 treatment? my data are non parametric

[@brensonyeo1358]

pakai Uji Kruskal Wallis yang ada di K independent sample test, jangan lupa input grouping variable kamu

[@madiharazzam1098]

i run this test and my output did not show the Test Statistics table. why is it so? it is only showing Ranks table.

[@rajnikantsharma]

In SPSS, Choose Analyze, then Descriptives, and then Explore Enter Continuous Variables in the "Dependent List" Column, and the Categorical Variable in the "grouping List" column. Choose your options under the following tabs: Statistics (Plots) (Options) Then hit OK to get the results

[@TheWolvesII]

Thanks for your video! Also good hint on the equality of non-normality assumption! Could you please explain the difference of U and Z? & then why, e.g. U can be positive and Z negative?

[@how2stats]

You'll find the answer to that question here: www.ibm.com/support/pages/mann-whitney-u-statistic-spss-differs-some-books-other-software-p-values-are-same

[@TheWolvesII]

@@how2stats thank you!

[@GGMacie]

Veri najs video :)

[@daisychain17]

What do the Z scores tend to mean?

[@dennisrosenberg389]

It tells you how many standard errors you are far from the mean of the population.

[@vanijain6525]

Oh god... thank uh so much ♥️

[@henokandualem3718]

it is good to use mean instead of median for non-parametric tests

[@ayaalialrubaie3714]

Thank you

[@mathkul3t]

thanks boss

[@benysmart1643]

Thank you too much

[@vrdlgmc-j2784]

thank you very much for such a nice lecture . Can i get you email id to discuss my statstic query with you.