I need to establish correlation coefficients between an independent ordinal variable (the scores are computed with the use of Likert scales) with an interval dependent variable (test scores). May I know do I use the Pearson r or the Spearman? Please advise. Thank you.
$\begingroup$
$\endgroup$
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
It is pretty common to consider Likert scales as quasi-metric, that is using them as if they were metric data is the heart of so called "Classical Test Theory". If the item answers were (rightfully) considered ordinal, you would not have been able to sum them up to compute the score in the first place. So the damage is already done, when you have computed sum scores or mean scores. In that case go ahead an use Pearson's, if you prefer to do so.
However, Spearman and Pearson correlation are not "somehow the same for different kind of data". See my answer here for examples where both give disturbingly different results: What is the explanation for having a Pearson's correlation coefficient significantly larger than the Spearman's rank correlation coefficient?
