Multiple dependant variables in logistic regression

Question

I have a binary dependent variable, say food security status, where 0 = food insecure and 1 = food secure, and this response variable is recorded for two categories of households, small farmers and land less rural house holds. Both these household categories may have food secure or food insecure. These dependant variables are a function of some common independant variables. I want to check which independant variable has more impact on farmers or landless households' food security.

Is it possible to include it in one logistic regression model? Or do I need to run separate models for farmers and landless households?

Answer 1

I don't think you have two dependent variables, I think you have one dependent variable: Food security. Then you have multiple independent variables, including whether the person owns land, and you want to look at the interactions of landowning and the other IVs.

Answer 2

I second Peter's answer. But if you really want to use both outcomes as dependent variables, then you may try a multinomial logistic model by transforming the two categories of food security and the two categories of land owner into 4 categories. Something like this:

Cat 1: Food secure and small farmer
Cat 2: Food secure and landless household
Cat 3: Food insecure and small farmer
Cat 4: Food insecure and landless household