1 For formal definitions see Equations (6.51), (6.69), and (6.70).
2 See Deaton (1997) for a full discussion of the issues involved.
3 For example, if the data are based on a simple survey of households, but one wants to infer something about the distribution of individuals, one needs to weight each household observation by an amount proportional to the number of persons in the household; this structure is similar to the weights introduced by stratification.
4 If, as is usual, trimming is something that is done voluntarily by the user rather than being imposed by the data provider, then Cases E and F are not relevant in practice.
5 A standard example of this “top-coding” of some income components in the Current Population Survey—observations ...