Effect size of a Kruskal-Wallis test

Discuss the jamovi platform, possible improvements, etc.

by RHainez » Tue Jun 04, 2019 8:41 am

Hi,

When running an anova, jamovi provides several effect sizes but non when we run a Kruskal-Wallis test. Do you think it's possible to add this feature ?
There is a discussion on ResearchGate.net about this subject with the explanations on how to achieve such a value:

https://www.researchgate.net/post/Anyon ... s_analysis

Have a nice day :)
RHainez
 
Posts: 57
Joined: Wed Feb 08, 2017 3:14 pm

by jonathon » Tue Jun 04, 2019 10:45 am

oh yup, i'll take a look.
User avatar
jonathon
 
Posts: 873
Joined: Fri Jan 27, 2017 10:04 am

by coledavis » Tue Jun 04, 2019 2:11 pm

epsilon squared (Tomczak and Tomcszak 2014).
k = kruskal.test(dv ~ iv, data = data)
k$statistic * (NROW(health) + 1)
/ (NROW(health) * NROW(health) - 1)
(used in my R book - https://vorpress.com/r/index.htm)
coledavis
 
Posts: 15
Joined: Sun Jul 02, 2017 1:35 pm

by coledavis » Tue Jun 04, 2019 4:07 pm

coledavis wrote:epsilon squared (Tomczak and Tomcszak 2014).
k = kruskal.test(dv ~ iv, data = data)
k$statistic * (NROW(dv) + 1)
/ (NROW(dv) * NROW(dv) - 1)
(used in my R book - https://vorpress.com/r/index.htm)
coledavis
 
Posts: 15
Joined: Sun Jul 02, 2017 1:35 pm

by jonathon » Tue Jun 11, 2019 9:03 am

cheers cole.
User avatar
jonathon
 
Posts: 873
Joined: Fri Jan 27, 2017 10:04 am

by coledavis » Tue Jun 11, 2019 1:08 pm

You're welcome. Other relevant material on this webpage:
https://www.researchgate.net/post/Anyon ... s_analysis

The original Tomczak and Tomczak article can be found here:
https://www.academia.edu/10618805/The_n ... ffect_size
coledavis
 
Posts: 15
Joined: Sun Jul 02, 2017 1:35 pm

by jonathon » Fri Jun 14, 2019 2:53 am

i've added this to the 1.0.1

cheers
User avatar
jonathon
 
Posts: 873
Joined: Fri Jan 27, 2017 10:04 am

by RHainez » Mon Jun 17, 2019 2:26 pm

Thank you very much, it works! :)
RHainez
 
Posts: 57
Joined: Wed Feb 08, 2017 3:14 pm


Return to General

cron