Closed joshyam-k closed 8 months ago
what needs to be done to allow bootstrap to still produce an mse estimate for a small area that we have no sample data for:
area_random_errors needs to include all domain levels from the population data and it currently just includes the ones from the sample data: https://github.com/harvard-ufds/saeczi/blob/6e468872de3c36aff7ad8b34800aa60e3e2ebacb/R/unit_zi.R#L116-L123
area_random_errors
joined_pop_bi$b_i is a vector which needs to be filled with 0 for all indices that belong to a domain not in the sample https://github.com/harvard-ufds/saeczi/blob/6e468872de3c36aff7ad8b34800aa60e3e2ebacb/R/unit_zi.R#L133
joined_pop_bi$b_i
other sampling functions may also need to be tweaked slightly
what needs to be done to allow bootstrap to still produce an mse estimate for a small area that we have no sample data for:
area_random_errors
needs to include all domain levels from the population data and it currently just includes the ones from the sample data: https://github.com/harvard-ufds/saeczi/blob/6e468872de3c36aff7ad8b34800aa60e3e2ebacb/R/unit_zi.R#L116-L123joined_pop_bi$b_i
is a vector which needs to be filled with 0 for all indices that belong to a domain not in the sample https://github.com/harvard-ufds/saeczi/blob/6e468872de3c36aff7ad8b34800aa60e3e2ebacb/R/unit_zi.R#L133other sampling functions may also need to be tweaked slightly