ApolloChoiceModelling forum

Hi David and Stephane,

I have a question about the VTT error calculated from a RRM model. If I understood correctly the paper of Daly et al. (2012) about the computation of this measure, we calculate errors from a transformation of the parameters, and in the case of VTT is the ratio of the derivatives of the deterministic parcel of regret/utility in relation to time and cost attributes. In RRM models the VTT is not simply the ratio between the travel time and cost parameters (Chorus 2012 paper) of the alternative considered. My question then is: does it make sense to use the apollo function apollo_deltaMethod to calculate the error of this transformation for a RRM model? Or should I compute it manually? Thanks in advance.


Best,

Hi

you definitely can't use that function in that case. But I also think you should be careful how you interpret RRM results and whether you really should refer to that calculation as VTT.

Stephane

Hi Prof Stephane

Thanks for the answer. I'm using as reference the Chorus (2012) https://www.tandfonline.com/doi/abs/10. ... 011.609947 and Dekker (2014) https://www.sciencedirect.com/science/a ... via%3Dihub papers. I'm aware of the differences between the RUM approach, but I've never seen a discussion about why shouldn't refer to the derivative of dRR/dT/dRR/dC as VTT. Can you indicate some reference about this? Thanks again.

Best,
Gabriel

Gabriel

in my view, VTT is a concept grounded in micro-economic theory, and one which requires consistency with utility maximisation. See https://doi.org/10.1007/s11238-017-9651-7

You can of course calculate a measure from RRM, but what would it mean, and how could it be used in economic analysis?

Stephane