Should't these be significantly negatively correlated?

#1
Hello!

I am wondering how these two variables don't come up as negatively correlated when using correlation matrix in jamovi.

These variables are scores for two different likert 1-to-5 scales.

For me, by eyeballing the individual data points of the two scales, it seems that they are clearely negatively correlated.



The individual average scale scores data points (to be eyeballed)

1652263393731.png


The results obtained after running the correlation matrix in jamovi (showing no significant negative correlation)

1652263925237.png


Am I doing anything wrong? Or the result obtained in the correlation matrix is normal?

Thank you!
 

Attachments

Last edited:

Karabiner

TS Contributor
#2
First of all, make an X-Y-scatteplot. Then, also do a Spearman rank correlation analysis,
in case that few extreme data points might have a huge impact on the (Pearson-)correlation.

With kind regards

Karabiner