Kruskall- Wallis or U-Mann Whitney

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

txema85
Posts: 2
Joined: Wed Dec 14, 2016 11:07 am

Kruskall- Wallis or U-Mann Whitney

Postby txema85 » Wed Dec 14, 2016 11:13 am

Good afternoon, I have a question regarding non-parametric tests and SPSS.

I have the concentration of a marker in blood and I want to see if there are differences with respect to the origin (I have several origins). On one hand I do the Kruskall-Wallis test, which gives me a significant p. If I enter the "model viewer", it gives me a comparison in pairs, with a significance and an adjusted significance, which do not resemble anything at all (the adjusted is always much higher, and many that are significant cease to be). On the other hand, if I do the Mann Whitney test pair by pair, I get a different p, similar to unadjusted significance. My question is, to establish if there are differences in concentration according to the origin, which p has to be chosen?

Thanks
statman
Administrator
Posts: 2756
Joined: Tue Jun 12, 2007 12:08 pm
Location: Florida, USA

Re: Kruskall- Wallis or U-Mann Whitney

Postby statman » Wed Dec 14, 2016 1:40 pm

2 different test so need more info.

Also ...........
See the note below

NOTE: Please read the Posting Guidelines and always tell us your OS, the SPSS version and information about your study and data!

Statman
Statistical Services
txema85
Posts: 2
Joined: Wed Dec 14, 2016 11:07 am

Re: Kruskall- Wallis or U-Mann Whitney

Postby txema85 » Thu Dec 15, 2016 8:51 am

First of all, thanks for answering so quickly. I use Windows 10 and SPSS v22.

About my study, I am studying a biomarker in cancer, and I want to know if there are differences related with the location of the cancer (head and neck, lung or cervix).
When I do the Kruskall-Wallis test, I find a statistically significant p. Deeper, If I click on the model viewer, I can see the comparison between pairs (head and neck and lung, head and neck and cervix, cervix and lung....) with a significance and an adjusted significance, which are very different.

On the other hand, if I decide to compare pairs using Mann-Whitney, I get a different significance.

My questions are:

Which are the different bettwen K-W significance and the adjusted one, in the comparison between pairs?
And, If I want to check if there are differences between the groups, head and neck and lung, head and neck and cervix, cervix and lung, which test should be performed?

Thanks again

Who is online

Users browsing this forum: No registered users and 1 guest

cron