Stats II: Multiple Regression Help

[Serif]

I know a lot of people on here are good with numbers seeing how they figure out all these formulas for pDIF cap or whatever. I am was wondering if anyone could help me on an extra credit problem with my Stats II.

We were given a 2 sets of data(male, female) that consisted of 1 dependent variable(salary) and 5 independent variables(age, years, rankhire, rank, yearsrank). We had a 6th independent variable but we removed it because of multicollinoarity.

Our initial project was to do either male or female, whichever was assigned to us, and figure out the best model for it. We used backward elimination for both the "Mix" Preference Algorithm and the "Outlier " Preference Algorithm. We then figured out the model for each and then based on the higher Adjusted R Square value we chose that model.

I already completed that model and I do not need any help with that part, my question is on the extra credit.

The extra credit problem is to calculate the amount of money that one needs to add to each female salary to bring the female salaries in line with the male salaries based on the ABSOLUTE BEST regression model that corresponds to your version and demonstrate that this amount is correct. Show all the steps used to calculate the amount and all the steps used to demonstrate its correctness.

Now my initial task was to figure out the best model for the males and now I am needing to figure out how to adjust the female model to make the line equal to the males.

I did the same test for both male and female and I got:

Males Best Model: Salary = 17, 951.22 + 3,941.71(Rank) + 394.72(YearsRank)
Adjusted R Square Value = 0.908955172

Females Best Model: Salary = 14, 749.99 + 89.97(Age) + 370.83(Years) + 1771.60(Rank)
Adjusted R Square Value = 0.930238873

I am going to ask my Professor today for some more help but I did not know if anyone could possibly give me some pointers or hints on how to figure out the extra credit part.

The only clue that he really gave us was that the only thing we will change is the dependent variable, salary.

[Creeps]

ugh, fuck stats

[Enkidu]

y=mx+b? So it's just the difference in the intersect plus the difference in the slope per independent unit?

[Serif]

I finally figured it out but I doubt anyone cares. If you do want to know how, send me a PM. I wish I could delete the topic though.