Mexican Family Life Survey
ENNVIH-1 (2002)
Weights
1. Weights Description
The weight is the value used to adjust the information of the variables captured through a sample in order to multiply the observations to better represent the whole population. These weights are unknown and depend on the specific sample design.
The basic weights are obtained from the inverse of the final probability of selecting the specific observation. We can obtain the correct weights by adjusting the probabilities of the basic ones, according to the final distribution of the sample.
There is an alternative way to adjust the weights, which is to use, when possible, the information available over the whole universe (population census, comprehensive surveys, etc). In this case, adjustments over the basic weights are just a matter of simple ratios so that the observations reproduce themselves to better represent the whole population.
In any of the two cases, it is very convenient to be aware of how these basic weights were constructed.
2. MxFLS-1 Weights
The design of the Mexican Family Life Survey sample is a very complex one: three-staged, by conglomerates and stratified (see MxFLS Sample Design, INEGI 2001) This is why the construction of the expansion weights is so useful. However, it turned out a very difficult task because, in order to adjust the basic weights, we had to consider the non-response rate of each book independently, such is the reason why each book has a different set of weights. Individually-based books (3A, 3B, 4, 5, EA, EN and S) will have and expansion weight at an individual level, whereas household-based books (C, I, II) will present weights at a household level.
3. Using the MxFLS-1 expansion weights
To use the expansion weights you need to merge the information you wish to analyze, with the corresponding file containing the related weights. Once having identified the book, the information should be merged using the variable "folio" for those household-based books, and with "folio" and "ls" for those individually-based books. After merging the data, you should always use the variable factor__# to assign the correct weight to each observation.
The files with expansion weights for each book of MxFLS-1 can be downloaded from the table below:
MxFLS-1 Expansion Weights
Book IIIA
Book IIIB
Book IV
Book V
Book Proxy
Book S
Book EA
Book EN