It appears that the coefficients for 3-person households and 4-person households are not being added to the right utility equations in estimation mode. These are the coefficients that are applied when the alternative is where all persons regardless of person type choose the same pattern. They are expressed as P_coeff_D_xxx, P_coeff_D_xxxx, and P_coeff_D_xxxxx, where D is either M,N, or H and the x’s refer to the number of persons in the household.
Here is the utility trace for alternative 1 (MMM) for a 3-person household. It looks like the all-household member interaction term for 5-person households ( P.coef_M_xxxxx) is being applied to this alternative instead of the interaction term for 3-person households ( P.coef_M_xxx).
As reported by @jfdman
It appears that the coefficients for 3-person households and 4-person households are not being added to the right utility equations in estimation mode. These are the coefficients that are applied when the alternative is where all persons regardless of person type choose the same pattern. They are expressed as P_coeff_D_xxx, P_coeff_D_xxxx, and P_coeff_D_xxxxx, where D is either M,N, or H and the x’s refer to the number of persons in the household.
Here is the utility trace for alternative 1 (MMM) for a 3-person household. It looks like the all-household member interaction term for 5-person households ( P.coef_M_xxxxx) is being applied to this alternative instead of the interaction term for 3-person households ( P.coef_M_xxx).