What statistical method should I use?

#1
I am trying to establish the impact of several different variables on our gender pay gap. I'm hoping to end up with a waterfall type graph showing each driver of the gap and the % impact that each variable has had on the change from 2018 to 2019. What analysis would you use to look into this if all variables are independent and don't relate to each other?

The data source contains a sample who earn varying amounts. The variables I wish to test include performance ranking, career level, upward feedback ranking and whether there was a reward for extensive travel.

Thank you for any help in advance.
 

hlsmith

Not a robit
#2
This seems like it could be examined via multiple linear regression. You will need to examine your distributions and model residuals - in that pat is usually a positively skewed variable. I don't work with waterfall plots, not sure if you standardize your predictors - that you could then generate one.
 
#3
Thank you for your reply. Assuming I can establish a strong positively skewed relationship for certain variables, is there then a way to estimate the % impact of each IV on the overall % change in our gender pay gap?