Excel - Friedman test

[@mohammedal-humaiqani5166]

@stikpit, thank you for the very informative video. Would be possible for you to share the excel sheet as well?

[@shamcontrol1460]

Hi, This is really useful. Thank you. Could you provide excel templates for post hoc tests after Friedman's as well (Dunn's test, Wilcoxon test, Sign test, Fisher LSD test). It will be of great help. Thank you.

[@stikpet]

Glad it helped. Unfortunately I don't have any proper template yet for this in Excel. I do have a work in progress, but made that a while ago not sure exactly what it is anymore, but perhaps it helps: bit.ly/2TXtP5q I do have some instructions for the post-hoc using R or SPSS. R is free and might actually be easier to do these things with then Excel. Those videos can be found on my site at: peterstatistics.com/CrashCourse/5-ThreeVarPair/ordinal/MultipleOrdinal3b.html Hope this helps

[@mimtodo9708]

The formula at 8.41 minutes doesn't work and it gives error in excel and it keeps highlighting the 0

[@stikpet]

hard to see what the problem might be, since in the video it works. Perhaps a minor typo?

[@aymanraouf8010]

Which video is the post-hoc test for the Friedman test?

[@stikpet]

Unfortunately I haven't yet created a video for any of the possible post-hoc tests for a Friedman test with Excel. What you could do for now is to use the Wilcoxon test kzbin.info/www/bejne/j2fWYWOPrrKnnNU and test each possible pair.

[@aymanraouf8010]

Not sure how to do the Wilcoxon test when I have 30 individuals all giving different rankings to a certain consequence type (5 variables)

[@stikpet]

@@aymanraouf8010 After you have done the Friedman Test and it has shown to be significant you can then do a pairwise Wilcoxon test. Take two of those five variables, and perform a Wilcoxon test with only those two (as shown in the video kzbin.info/www/bejne/j2fWYWOPrrKnnNU), that will inform you if those two are significantly different. Then take two other variables of those 5 and do the same. Keep doing this until you have done all possible combinations.

[@karencorredor]

Hi, the test that you runned is one way? I'm trying to figure out how to run Friedman for a randomized block design? could you give me a little help? I really appreciate your videos!!

[@stikpet]

To my knowledge the Friedman test is always for a randomized block design, but instead of blocks and treatments I use respondents and variables. See for example wps.prenhall.com/wps/media/objects/11886/12171343/OnlineTopics/bbs12e_onlinetopic_ch12-9.pdf or www.itl.nist.gov/div898/software/dataplot/refman1/auxillar/friedman.htm Hope this clarifies it.

[@aymanraouf8010]

Are you able to provide a template of this? Many thanks!

[@stikpet]

Hopefully today I can upload the example file. It might actually also contain some post-hoc tests but will need to check.

[@aymanraouf8010]

@@stikpet Looking forward! Many thanks for your videos again!

[@stikpet]

@@aymanraouf8010 Here's the example file used in the video: bit.ly/2TXtP5q. It also has some tabs for post-hoc. Note that this is still a rough file, I'd need to clean it up to be used as a nice template, but that will take a long time. Hope it still helps.

[@JackPerrin97]

What does this test show?

[@stikpet]

I briefly try to explain this on the website at: peterstatistics.com/CrashCourse/5-ThreeVarPair/ordinal/MultipleOrdinal3a.html does t hat answer your question?

[@JackPerrin97]

@@stikpet thanks