What correlation should I run?

[sonishell]

I have two data sets from identical ordinal scales (Never, Rarely, Occasionally, Often, Very Often). What correlation should I run?

 

[hlsmith]

Spearman Rank Correlation would work.

 

[noetsi]

Or possibly polychoric correlations if your software supports that and you feel there is a latent variable behind the ordinal measurements you are using.

 

[hlsmith]

noetsi,

Please explain this statement in more detail, " latent variable behind the ordinal measurements". Why might that need to be present?

 

[noetsi]

Because polychoric correlations assumes that behind the observed measurement there is a continuous latent factor. If there is not it is not valid to use polychoric/tetronic correlations (although in practice I suspect that is ignored).