mixed logit model : Interaction with socio-demographic variables

Hi

I am using stata user written command mixlogit in stata.

I have three alternative specific variables namely parking search time, park cost, walk time, (SP data set)

Socio-demographic attributes: Gender, Age group, Work Status (Created dummy for these categories)

I have interacted each alternative-specific variables with gender, age group, work status.

Outcome with interacted models: All most all are statistically insignificant.

My questions:
1. Is interacting socio-demographic with alternative-specific variables a common practice ?

2. Can I try in stata to add either log-normal or triangular distribution for alternative-specific variables and does the interacted variables require to follow their respective alternative-specific variables distribution ?

Your feedback would be greatly appreciated.

Thank you.

Regards
Alex
Sydney, Australia