Splitting nominal variable / correlation

Discuss statistics related things
Post Reply
grk99
Posts: 8
Joined: Tue Aug 09, 2022 5:03 pm

Splitting nominal variable / correlation

Post by grk99 »

Hi,

I want to carry out a correlation between pre/full-term and several dependent variables (ADHD, self-esteem etc..) ... To identify whether birth status affects different outcomes in adulthood.

How can I split the pre/full-term category into two separate variables please?

At the moment if I run the whole variable through a correlation, it doesn't split them, and only gives me one statistic.

Thanks
Attachments
Screenshot (64).png
Screenshot (64).png (123.04 KiB) Viewed 4372 times
grk99
Posts: 8
Joined: Tue Aug 09, 2022 5:03 pm

Re: Splitting nominal variable / correlation

Post by grk99 »

But I also know you can't feed nominal data through a correlation, so would I just compare means instead?
Attachments
Screenshot (66).png
Screenshot (66).png (100.56 KiB) Viewed 4369 times
User avatar
jonathon
Posts: 2613
Joined: Fri Jan 27, 2017 10:04 am

Re: Splitting nominal variable / correlation

Post by jonathon »

you could insert two computed columns, using the formulas:

IF(`Pre/Full Term` == 'Pre-term (before 37 weeks)', ADHD)
IF(`Pre/Full Term` == 'Full-term (after 37 weeks)', ADHD)

ideally jamovi would provide a 'split' function, but we haven't implemented that yet.

cheers
simonmoon
Posts: 17
Joined: Thu Oct 27, 2022 3:45 pm

Re: Splitting nominal variable / correlation

Post by simonmoon »

I think I replied to your other question about error messages in jamovi. I believe your issue is not the program related but more conceptual, measurement issue related. MacCallum et al.'s (2002) article in Psychological Methods will help and answer your questions.
Post Reply