Weighting when moving to subregion

[samclifford]

https://github.com/samclifford/MeganVermaMsc/blob/6e98e78cb50fad0a9a3b493f87f7845eec2c95f8/mscprojectscript_03_explore_data.R#L92

You probably need to look at the national population rather than the study size, as if we have lots of studies from a small country then that won't be as representative of the entire subregion.

[samclifford]

The wpp2019 package should be enough here

[mverma29]

what year should i draw the population from, given that we're summarizing here? just the most recent but for the entire subregions?

[samclifford]

Using each country's most recent is probably the best approach. I think wpp2019 has 2020 data in it.

[mverma29]

so i should first get a weighted average for each country based on total tested in each country (over the studies), then take the country weighted average to the subregions, where i will be weighting by total pop?

[samclifford]

Use the country level weighted averages that are already calculated. Merge with the country's latest population from wpp2019 (or WDI if necessary). Group by UN subregion and calculate weighted means.

[mverma29]

fixed in e84b0ec