|
Stata understands four kinds of weights:
1. fweights, or frequency weights, indicate duplicated observations. fweights are always integers. If the fweight associated with an observation is 5, that means there are really 5 such observations, each identical. 数据集中该样本被重复观测的次数
2. pweights, or sampling weights, denote the inverse of the probability that this observation is included in the sample because of the sampling design. A pweight of 100, for instance, indicates that this observation is representative of 100 subjects in the underlying population. 数据集中，该样本所能够代表的潜在分总体（sub-population）数。
The scale of these weights does not matter in terms of estimated parameters and standard errors, except when estimating totals and computing finite-population corrections with the svy commands; see [SVY] survey.
3. aweights, or analytic weights, are inversely proportional to the variance of an observation; that is, the variance of the jth observation is assumed to be σ 2/wj , where wj are the weights. Typically, the observations represent averages, and the weights are the number of elements that gave rise to the average. For most Stata commands, the recorded scale of aweights is irrelevant; Stata internally rescales them to sum to N, the number of observations in your data, when it uses them. 数据集中，该样本来源的观测样本数。
4. iweights, or importance weights, indicate the relative “importance” of the observation. They
have no formal statistical definition; this is a catch-all category. Any command that supports
iweights will define how they are treated. They are usually intended for use by programmers
who want to produce a certain computation. 数据集中，该样本的重要程度。
| 
发帖
雷达卡
京公网安备 11010802022788号
