Hello Sidra,
You could:
1. Use ordinary least squares regression, but if the IV had more than two levels or values, you'd have to recast it into k - 1 dichotomous variates (e.g., through dummy coding, or some related method), where k is the total number of levels for the ordinal IV. For a two-level IV, just code the values as 0 and 1 (functionally, this becomes a test of whether the point-biserial correlation is equal to zero).
Of course, it would be good practice to verify that OLS assumptions were reasonably satisfied!
2. Use ordinal regression.
3. Figure a way to scale the ordinal variable (I've included a couple of links below of people scaling ordered things), then use OLS regression.
4. Use one of the hybrid methods proposed by either Puri & Sen, or Iman & Conover. The latter is a tiny bit simpler to apply.
Good luck with your work.
Scaling adverbs: https://aclanthology.org/R15-1071.pdf
Rasch approach to scaling ordinal observations: https://www.rasch.org/memo44.htm
Puri & Sen method: Puri, M. L., & Sen, P. K. (1971). Non-parametric methods in multivariate analysis. New York, NY: Wiley.
Iman & Conover method: Conover, W. J., & Iman, R. L. (1980). Rank transformations as a bridge between parametric and nonparametric statistics. The American Statistician, 35. 124-129.
Sidra Cheema This question of yours is very general and depends on the assumptions and ideas of your study and research. However, I definitely recommend that you read David Morse answer carefully.
In any case a Spearman correlation coefficient tells you the basic information
- Badges
- Science method