Bivariate correlation analysis should be wrong. If the independent variable is an ordered set of classified data, it should belong to multivariate statistical analysis, and multiple regression analysis can be considered. When writing a thesis, I suggest that it is best to do factor analysis first in data analysis. You can describe independent variables into several categories, and then do multiple regression analysis to discuss the relationship between independent variables and dependent variables.

Of course, if the programming ability allows, machine learning models, such as XGBoost and random forest, can be considered, which should have good accuracy for this problem.