I have a statistics course assignment regarding the "pure" effect of the mileage on second hand cars' sales price.
The dataset contains several factors which may affect the sales price of cars on an exchange website, including:
- Year manufactured
- Mileage
- Make
- Type (Sedan, Wagon, SUV, etc)
- Color
- Complete logbook service (Y/N)
- Fuel efficiency
- Seller Zip Code
My understanding of the analysis of the "pure" effect of one independent variable on the dependent variable should limit all the other variables as the same, as in same make, manufacturing year range, type, color, etc. However, if I do that just for a single combination of cars of the same characteristics, I'd give up many data points.
So what's the best approach to tackle this kind of problem? Should I do many sets of single-variable linear regressions between mileage and sales price on many combinations of similar cars and average the effect?
Sorry there isn't any data here. I just want to have a road map of solving the problem. Thanks very much.