I'm currently trying to select the appropriate significance test and test of association and would really appreciate some help.

My independent variable is an ordinal subjective risk assessment (1-4) and the dependent is continuous (amount of data shared). I have yet to determine if I can use parametric stats, but as of yet the distribution does not appear normal, nor is variance homogeneous.
I would like to determine the strength of association between risk assessment and data shared. I would also like to determine whether or not those in each risk assessment category vary from one another. I therefore likely won't be able too use anova.

I had thought about using a chi-square, but chi-square does not take into account the rank. I also don't want to split my dependent variable into categories if I don't have to. I had also considered Spearman's Rho for the strength of association.

Am I thinking about this all wrong? Any help would be greatly appreciated.
What is the correct statistical technique when using ordinal predictor variables and continuous outcome variables.