Survey Weights in the Construction of Adult Literacy Rates

[Chris Hickey]

Hi,

I have household survey data from a country with the following variables:

district ID (distid)

household ID (hhid)

size of household (nrt)

total household monthly expenditure (expr)

educational attainment (in years)

A dummy variable for whether the individual is literate (0) or not (1)

Age of Individual

household weight

I am wanting to calculate the

Adult Literacy Rate

for each district in the sample. That is, I want to calculate the % of the population aged 15 and above that is literate. However, I am unsure of how to incorporate the household weights into this calculation and what stata command to use.

I have already used the weights to calculate average monthly expenditure per capita in each district:

gen exppc = expr/nrt
collapse (mean) exppc [w=wert], by (distid)

And I am comfortable with that application and result. However, I am uncertain of how to make use of the household weights when calculating the indicator for literacy rates.

Any advice would be appreciated.
Thanks

[Chris Hickey]

I have tried the following command but the result I get for literacy rates is the same as if I run the command without the weights.

gen indiv = 1
bysort distid: egen adpop = sum(indiv) if age >= 15
collapse (sum) indiv if illit < 1 & age >= 15 [w=wert], by (distid provid adpop)
gen lit=indiv/adpop

[Rich Goldstein]

your question is not very clear to me and I don't see anything about how you have svyset your data either; the following might help

help svyset
help svy_estimation

in the second of the above help files, pay particular attention to the "descriptive statistics" section