Weighting efficiency provides the amount of distortion needed to arrive at the weighted figures - i.e. how much the data is manipulated by the weighting.
The result is stored in the weighting report located in the survey.dat directory.
A low efficiency indicates a larger bias introduced by the weights.
Here is the formula that is used to calculate it:
|w i||w i²|
|WE||= 616.080273 / (24 * 32.0645763)|
The weighting reduces the reliability of the sample of 19.1% (~ 100 - 80.06 ~)
In other words, it's as if we have removed 5 people from our sample of 24
(24 X 80.06 % =19)
Our sample is matching to 80.06% of the population and the effective base is equal to 19
When the weighting efficiency is below to 80%, this can mean that you have a mismatch between the sample and the population.
And when WE is below 70%, we recommend you to check the weighting design or perhaps even go back to the sample design.