Using Weights for countries in Survey Dataset

[Anna Petrova]

Hello everyone,

I am working with 2 waves (5 and 6) of World Value Survey in Stata 15 and I am analyzing the effect of different individual characteristics as well as different generational groups on participation in charity.

I have grouped countries into 4 groups by cultural characteristics: each group contains 5-6 countries


I am using logistic model with wave fixed effect and country RE

melogit charity_participation i.generation sex i.income_class i.educ_level happiness char_conf marital_status i.religion i.wave || country:

I have read some papers which advise to use weight variables. In WVS Longidunal Dataset there is a variable S018 and S019 which are both factors derived to transform the sample's N to 1000 and 1500.

1) Can you advise if I need to use these weighted variables in my analysis. And, if yes, how I can generate the population based weight in STATA (coding).

There is an example on WVS website in SPSS style:

COMPUTE POPWEIGHT=0.

FORMAT POPWEIGHT (F15.5).

IF (S003=8) POPWEIGHT=S018/1000*3639459. /*Albania*/ where 3639459 - number of population S003 - country variable

IF (S003=12) POPWEIGHT=S018/1000*33769669. /*Algeria*/

2) If I create the weights for each of the country where I need to use it after in my regression.

Thanks in advance.

BR