This is the repository for the English version of the Future Elderly Model, originally developed at the Leonard D. Schaeffer Center for Health Policy and Microsimulation.
Found lnly values of 0 which should not be possible in the model. Have traced these values all the way back to the long file, where I think the 0 values come from missing values, most of which are due to the respondent not being in the self-completion sample.
The mystery here then is why the kludge.do file is not properly imputing these, despite going through a hotdecking step and filling current from lag and vice versa. Need to check the hotdecking step and figure out whats going on, potentially fill the still missing values with 1 as that is most common by far (over 55% of stock pop).
ld-archer
commented
1 year ago
Found lnly values of 0 which should not be possible in the model. Have traced these values all the way back to the long file, where I think the 0 values come from missing values, most of which are due to the respondent not being in the self-completion sample.
The mystery here then is why the kludge.do file is not properly imputing these, despite going through a hotdecking step and filling current from lag and vice versa. Need to check the hotdecking step and figure out whats going on, potentially fill the still missing values with 1 as that is most common by far (over 55% of stock pop).