Hello!
I work on the Brazilian 1980 census.
I have already asked a question about the household weight variable WTHH having some values equal to zero. Tim Morelan replied to me that it was because household questions were asked only of those living in “permanent private households.” So some individuals are included in the person-level analysis but not in household-level one. This makes sense.
My question is now about the individual weight variable WTPER. I have noticed that 10 individuals had a value zero for WTPER. This does not make sense to me. Can you tell me the reason why these individuals are excluded from the person-level analysis?
Thanks a lot for your help.
Geoffrey
It may be helpful to refer to the sample design for the 1980 Brazil Census, specifically the section on sample weights. Brazil’s statistical agency employed a post-stratification process based on age, sex, urban/rural, and other characteristics. Each strata was then assigned a weight, and each person was assigned an integer weight to make the final weighted sample representative of the universe. This suggests that a zero weight indicates a person included in an overrepresented stratum.
However, as the weight values were created by the Brazilian Institute of Geography and Statistics, there may be an alternative explanation for the presence of zero weights. Thus, contacting them directly may be the best method of determining why these 10 cases were assigned weights of zero. Keep in mind that such a small number of cases is unlikely to affect your analysis.
Hope this helps.