ApolloChoiceModelling forum

Hello,

In my dataset, I have many socioeconomic variables such as age_group, Disability, Income_group.

I have 4 mode choices including Drive, Ride, Bus, Walk.

When I put all the socioeconomic variables in each of the utility functions, it leads to over-specification of variables. What is the solution to this?
Is it correct if I put a socio-economic variable in just one or two utilities instead of all four utilities?

V[["DR"]] = asc_DR + A_Drive_Time * Drive_Time + A_Drive_Cost*Drive_Cost
V[["RI"]] = A_Ride_Time *Ride_Time
V[["WA"]] = asc_WA+ C_Walk_Time*Walk_Time +C_Age*age_group+C_Disability*Disability
V[["BUS"]]=asc_BUS+ D_Bus_Time*Bus_Time +D_Bus_Cost * Bus_Cost +
D_Income_Group*(Income_Group==0)+ +D_Age*age_group+C_Disability*Disability

Thanks!
Ariana

Ariana

please remember that only differences in utility matter. So if you just interact the socio-demographics with the constants (which is what you do below), then you can only include them in J-1 alternatives. It's different of course if you interacted them with the continuous attributes

Stephane

Hi Stephane,

Thank you so much for the clarification. I really appreciate your help!

Best,
Ariana