search

athowes / multi-agyw

Spatio-temporal estimates of HIV risk group proportions for AGYW across 13 priority countries in sub-Saharan Africa
https://athowes.github.io/multi-agyw
MIT License
2 stars 2 forks source link

Inclusion of Kish effective sample size as offset #22

Closed athowes closed 3 years ago

athowes commented 3 years ago

Might improve the recovery of the Poisson trick sample size (i.e. reduce the standard errors, since the mean seems right), as Tim found in his simulations. That said, there doesn't appear to be a problem with the model so perhaps not required

athowes commented 3 years ago

Tried doing this using E = n_eff_kish in the R-INLA formula. The recovery looked pretty visually similar, could be investigated further maybe

athowes commented 3 years ago

image

athowes commented 3 years ago

Perhaps it's nice statistically, but doesn't seem that important in that it would materially effect the results, so going to close this for now. Perhaps come back to another time