# Log-Log Multiple Regression - How to incorporate discount + interpreting

#### nils88

##### New Member
Hello smart people!

I am working on a project where I’m trying to understand consumers’ price elasticity, and I was hoping I could get some input from you statistics/econometrics experts (my skills are not very advanced…). I am using historical sales data in order to assess the impact of pricing and promotional decisions. I have developed a log log regression model in order to do so. I have also attached a sample of my doc!
My main question is if I should use:

1. Ln Average sales per salesman = a + b1*Ln Discount Price + b2*On Discount Dummy + b3*On Display Dummy

Or

2. Ln Average sales per salesman = a + b1*Ln Standard Price + b2*Discount In % + b3*On Display Dummy

In other words, if I should use: Ln Discount price + On discount dummy OR Ln Standard Price + Discount in %.
I have not done a multiple regression analysis with discount before so I am not sure on how I should incorporate it into my model. Also, in the second model, lets say b2 is 7,26, how do I interpret it?

Another thing is that I am a bit concerned about the very high b1. That a 1% increase in price would lead to a -3.58% change in demand. It does not feel in line with reality but perhaps it is. Any comments based on what you can see in the attached .xls?