I have been exploring the assembly line approach to estimating CDEs. Given the data I am working with, I want to incorporate weights into my outcome regression. I have tried to do this by passing , weights = my_weights to the end of the out_reg() function. I do this for both the treatment and the mediator. However, the error message Error in model.frame.default(formula = .de_y ~ ... : invalid type (language) for variable '(weights)' is returned. I am using the logit engine to estimate the propensity score and the lm engine to estimate the outcome regression. Is it not possible to incorporate weights into the outcome regression at this time? If it is, in which portion of the assembly line does one incorporate them?
I have been exploring the assembly line approach to estimating CDEs. Given the data I am working with, I want to incorporate weights into my outcome regression. I have tried to do this by passing
, weights = my_weightsto the end of theout_reg()function. I do this for both the treatment and the mediator. However, the error messageError in model.frame.default(formula = .de_y ~ ... : invalid type (language) for variable '(weights)'is returned. I am using the logit engine to estimate the propensity score and the lm engine to estimate the outcome regression. Is it not possible to incorporate weights into the outcome regression at this time? If it is, in which portion of the assembly line does one incorporate them?