isidro survey phone call

[aspina7]

to be finalised once dictionaries are available:

• [ ] add a note to explain why age_group and sex are included as strata (that would normally be the sampling strata) and also that correcting for this is difficult in situations where dont have an accurate population estimates
• [ ] add commented-out code for generic weight calculations so ppl can account for population differences in clusters (isidro to send alex)

other points discussed:

• isidro to ask con about finite population correction
• tab_* broken -> gtsummary wrapper
• melbourne {epiR} -> sample size - potential use

[pbkeating]

The current add_weights_cluster() function for survey analysis is only applicable to surveys that don't involve PPS. A standard two stage cluster survey should have cluster-specific (e.g. village specific) weights while a two stage cluster survey with PPS will have the same weight applied to all clusters (see attached doc). The formula behind the current function doesn't take into consideration whether PPS was applied or not.

The formula to use with a two stage cluster survey that employed PPS can be found in the attached document

psws_probability_prop_size_bierrenbach (1).pdf