search

epix-project / mics

Utilities to analyse multiple indicator cluster surveys
Other
1 stars 0 forks source link

Create second weight variable in transform step to denormalise data and estimate numbers in strata #13

Open soniwe opened 6 years ago

soniwe commented 6 years ago

  1. Create a new column in the transformed dataframes by merging with worldpop data for women aged 15-49 for the related country and same year as survey.

  2. Create a new column in the transformed dataframes for number of women surveyed (need to extract this from the survey reports or meta-data)

  3. Calculate new norm_weight =

Example Stata code for Kenya 2014: *Rescaling of the sampling weight to sum to the population estimate for Women 15-49

generate pop_women_15_49 = 10549000 /UN population (world prospect) estimate of women 15-49 yrs old in Kenya in 2013 /

generate weightpop = pop_women_15_49v005/(310791000000) / 31,079 women responded to the survey /

*Define the sample design (DHS VII) svyset [pweight=weightpop], psu(v021) strata(v022)