1

Could someone point me toward a specific method to model data that consists of two groups of observations having the same dependent variable and sharing some explanatory variables, BUT also having explanatory variables that are defined for one group and not for another? A situation like this:

  • binary dependent var: y
  • shared explanatory variables for all groups: x1 and x2
  • if group=1, then x3 and x4 are among explanatory variables
  • if group=2, then x5 and x6 are among explanatory variables

My first reaction was to interact the group variable with the non-shared variables, but I don't think that's a good idea.

I appreciate any hints and clues.

Moe
  • 141
  • 7
  • The closest thing I've found so far is a structure similar to a SUR model. – Moe Aug 12 '15 at 06:07
  • A good answer will really demand some more context, you should explain us why this strange structure occurs. But, one similar example is when groups are men/women and only women have data about pregnancies. For such cases, see https://stats.stackexchange.com/questions/372257/how-do-you-deal-with-nested-variables-in-a-regression-model/372258#372258 If your case is different from this, tell us your context. – kjetil b halvorsen Feb 24 '20 at 01:36

0 Answers0