Nonparametric test - which test should I use?

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

DIversity
Posts: 2
Joined: Mon Sep 22, 2014 12:32 pm

Nonparametric test - which test should I use?

Postby DIversity » Mon Sep 22, 2014 12:37 pm

Hello!

I want to compare two groups: green-wearers versus orange-wearers (DV)

I have 20 subjects: 10 wore green, the other 10 orange.

One variable is how dominant these subjects were (classified 1 to 7) = real-world data.

The other variable is the mean (not nominal data!) of how dominant other participants (another experiment) rated subject 1-20 were.

E.g., I know subject 1 is "average dominant" (classified in "4"). The total mean he received from others (N=211) on how dominant they thought he was is "3,88".

Now, how can I compare this? My hypothesis is that orange wearers were perceived to be more dominant than green wearers compared to how dominant they actually are.

Does this make sense?
Thanks a lot!

P.S.: I am using SPSS 20 ... and I checked under "HELP" before I posted this.
statman
Administrator
Posts: 2700
Joined: Tue Jun 12, 2007 12:08 pm
Location: Florida, USA

Re: Nonparametric test - which test should I use?

Postby statman » Mon Sep 22, 2014 12:44 pm

Look into one of the Independence tests

Read the Help and check Case studies

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
DIversity
Posts: 2
Joined: Mon Sep 22, 2014 12:32 pm

Re: Nonparametric test - which test should I use?

Postby DIversity » Tue Sep 30, 2014 7:13 am

Here's me again with my orange-green-wearer-problem:

My hypothesis is that orange wearer’s actual handgrip-strength will correlate more positively with how dominant they were perceived than it will within green wearers.

As above ... e.g., I know subject 1 (out of 20) is "above average strong" (classified in "6"). The total mean he received from others on how dominant they thought he was is "3,88", he wore green.

Since a correlation isn't possible, and multiple lin. regression neither (sample size! I had 211 participants rate 20 images ... N=20 is too low), how can I find out whether there is a pattern? I.e., the stronger someone was, the higher the ratings for dominance he received ... and even more so when he wore orange (rather than green).

I assume crosstabs is the best way to go. But I am not sure what's the best way to arrange things. Any suggestions please?

Who is online

Users browsing this forum: No registered users and 2 guests

cron