Plots to show between country level variance as compared to within country

[athowes]

• Similar to page five from here
• Could colour by age

[athowes]

Things to improve from first version:

• Sort out the axis with DHS survey names. Want ISO3 on the outside and then just the year on the inside. Ideally would have the surveys from same country grouped closer together
• Perhaps having the time shown here isn't worth it and better just to filter down to most recent survey and show that, not sure what story the time is telling
• Add country-level mean (simple: just mean, little harder: population weighted) as white dot, similar to example
• Have 0%-100% rather than 0.00 to 1.00
• Maybe switch facet grid around to have age on the top
• It's going to be very busy with 13 countries

[athowes]

Linking to #20: one reason to fit all of the countries jointly rather than individually is that it would allow quantification of the level of between country variance via a country random effect. This can be shown with this plot, but it isn't a quantification.

[athowes]

Relatively happy with this. Will close until coauthor review.