search

ld-archer / E_FEM

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.
MIT License
3 stars 1 forks source link

Add dummy variables for different types of poverty #74

Open ld-archer opened 3 years ago

ld-archer commented 3 years ago

Add some binary indicator variables for different types of poverty (wealth and income based).

Details to follow, but wealth poverty is defined in UK as having wealth below 60% of median value in a population.

This will need to come after the issues with wealth and income variables mentioned in #70 are resolved.

ld-archer commented 3 years ago

Using definition of poverty from here (UK gov).

They define income poverty as having income below 60% of the median in a population for a given year, the population in our case being the long format of ELSA.

ld-archer commented 2 years ago

Having a rethink about this. Decided to split the population into deciles by wealth and poverty, and will instead transition between these levels instead of a single binary poverty indicator (still have the poverty indicators and can compare later). Can use dummys for the poorest and richest groups to see how prediction is doing (can't use this in ROC validation as I think ROC needs absorbing binarys, won't work if respondent can move in and out of a group).