ApolloChoiceModelling forum

Hi Apollo community,

I'm a rookie for choice model and Apollo package.
I'm currently estimating an MNL (multinomial logit) model for a base model and have a few questions regarding the use of the correlation matrix in the output:

1. Should I opt for a classical correlation matrix, or would a robust correlation matrix be more appropriate?
2. Is it appropriate to use the correlation matrix from the output to check for correlations between coefficients?
Additionally, the output of the correlation matrix indicates a high correlation (over 0.6) between certain coefficients, such as b_price and b_range. It also shows a high correlation between each level of dummy-coded coefficients.

Given this case, is it acceptable to estimate the model by selecting either b_price or b_range, instead of including both?

Thank you in advance for your advice on these points.

Hi

you would find answers to these queries in standard textbooks on choice modelling. I would use the robust covariance matrix. 0.6 is not an issue - inspecting the covariance matrix is useful to see if there are any nearly perfectly correlated parameters, but 0.6 is way off that

Stephane