All,
I am working with a choice model for travel modes, where two of the modes - cars and Ubers - are incredibly similar to the point that their travel times are almost entirely correlated in the real world (as those modes would be using the same roads and be subject to the same road conditions).
Estimating the model itself via cmclogit was no problem, however, I am struggling with how to specify the margins command to test what would happen in a scenario. I am using the following command to see what mode shares the model would predict in a scenario in which the travel times of Uber and of taking a car would each increase by 10%:
Is there a post-estimation command that allows me to test what would happen if both alternatives' travel times increased by 10% simultaneously (while none of the other alternatives' travel times would change)?
Thank you!
I am working with a choice model for travel modes, where two of the modes - cars and Ubers - are incredibly similar to the point that their travel times are almost entirely correlated in the real world (as those modes would be using the same roads and be subject to the same road conditions).
Estimating the model itself via cmclogit was no problem, however, I am struggling with how to specify the margins command to test what would happen in a scenario. I am using the following command to see what mode shares the model would predict in a scenario in which the travel times of Uber and of taking a car would each increase by 10%:
Code:
margins, at(time=generate(time)) at(time=generate(time*1.1)) alternative(car uber)
Thank you!
