Wilcoxon Signed-Rank Test in SPSS with Effect Size Calculation in Excel

This video demonstrates how to conduct a Wilcoxon Signed-Rank Test in SPSS and the corresponding effect size calculation in Excel. The Wilcoxon Signed-Rank Test is the non-parametric equivalent to a paired-samples t-test.

[@dianaquirozhernandez8057]

Gracias desde México, no había encontrado información sobre este método y aun cuando mi ingles es deficiente lo entendí perfectamente, salvaste mi tesis. GRACIAS

[@roxana9483]

Thanks very much for this.! Do you have a reference you can share for how you determine less than -0.3 is moderate and less than -0.5 is large? I have seen a lot of examples for interpretation of Cohen's d for positive values but not for when they are negative with this formula. Thanks so much!

[@pratibhameena3546]

can you please answer this? thank you!

[@welxadnum]

No reply...

[@sarahburrous7374]

I liked the comparison table of the pre vs the post test. I will need to watch again to understand the results. If 65 out of 100 had no change, to me that means the technique didn't really work- even though the null hypotheses was rejected because the positive was higher than the negative.

[@strejdaMarty]

Saved my ars once again, dear Todd!

[@afidai]

Thanks for the very informative video. What about converting the z value to d value in calculating 95% confidence interval for that d value. As you know most statistical packages require that you have the d value along with the standard error or the confidence interval to conduct a meta-analysis. I don't want to exclude the studies from my meta-analysis that provided the u-value instead of something that I could use directly into a meta-analysis. Thanks in advance for your help.

[@CarlaWarneke]

Thank you Dr. Grande. How would I go about calculating the confidence interval for this effect size r = Z/sqrt(N)?

[@lowbird123123]

Dear Dr. Grande, is it possible for you to release a Tutorial about SPSS calculation of a minimally clinically important difference (MCID) with an anchor based method? Its hard to find something on the Internet. Your Videos are great and helpful! Thanks!

[@LadyDeLane]

Thank you! Great video, helped me a lot!

[@ahmedaltimimi9697]

I was wondering if you have any further information or papers i can quote on the cut off values of what is large moderate and low effect size. this would be greatly appreciated

[@soniasharmin4488]

Hi Todd, would you please answer this question?

[@hm.tusher]

SUPERB!!!!!!!!!!!!!!!!

[@aprilbrooks8672]

My understanding is that we will run this test only in cases where we can not run a non parametric test because it does not meet the data requirements.

[@mohammadhmodi5323]

Thanks you are wonderful

[@amandachaffin6636]

I liked how the Wilcoxon Signed-Rank Test compared the results to the median saying if it were greater/less than. I am not understanding what a moderate effect size means? How does that affect the results?

[@cassieperoulis4881]

+Amanda Sutton, I felt the same way and had similar questions. The link that Laurel Markus listed does seem to help some!

[@VishavJyoti-lq1kc]

Why Z divided by 200 not 100?

[@soniasharmin4488]

Hi Todd, Please answer this question.

[@frenky89]

@@soniasharmin4488 100 participants × 2 measurements = 200 data point = N (it is not the number of pairs but the number of data points)

[@bartul5645]

Which one is the true?

[@timstauffer1351]

Thank you.

[@bessidhoummahfoudh3757]

as for N...do we take all sample (control + exp) for example 16+16 =32 or should we take just one for the test we are doing. Fo example when doing Wilcocoxon for just experimental group, should N be like 16 and the same thing for control? Thank you

[@bartul5645]

did y learn?

[@wanderlustwernweh5313]

​@@bartul5645wot learn ?

[@ollybash2287]

why is n 200?

[@nvisscher]

In the Mann-Whitney video you went a step further and calculated the r squared... why is that not done here?

[@richardnightingale6371]

I would also like to know that

[@noreennudalo]

where can i find data with more than 30 samples??please??thank you

[@lirachannel1]

if the formula says N why do you take 2N? In your case it should be 100 and not 200.

[@welxadnum]

I think that's for 100 pretest and 100 posttest.

[@mariadt4124]

Too silent voice - I needed to put my volume to maximum to recognise.