Hello,
A model I work with is responsible for the allocations of funding for healthcare resources. Traditionally this has been a linear cost model. However, sometimes the ethnicity coefficients in the model can be negative, and a policy decision has been made that ethnicity can not be the basis for reduction of allocation resources. In these circumstances the coefficients are set to 0, and the predictions are run. However, in a new model, we use areg and predict using the option xbd. When I run some code recall the matrix of coefficients, and set some of the coefficients to 0 after areg has ran, I use predict yhat, xbd , however, Stata gives an error message: option xbd not allowed.
My question is whether there is anyway round this i.e. can I recall the the matrix e(b) make the changes necessary and then get the predictions using the xbd option.
N.B. I understand that setting coefficients 0 is not typical or usual but it is a policy decision I need to work with.
A model I work with is responsible for the allocations of funding for healthcare resources. Traditionally this has been a linear cost model. However, sometimes the ethnicity coefficients in the model can be negative, and a policy decision has been made that ethnicity can not be the basis for reduction of allocation resources. In these circumstances the coefficients are set to 0, and the predictions are run. However, in a new model, we use areg and predict using the option xbd. When I run some code recall the matrix of coefficients, and set some of the coefficients to 0 after areg has ran, I use predict yhat, xbd , however, Stata gives an error message: option xbd not allowed.
My question is whether there is anyway round this i.e. can I recall the the matrix e(b) make the changes necessary and then get the predictions using the xbd option.
N.B. I understand that setting coefficients 0 is not typical or usual but it is a policy decision I need to work with.
Comment