|
⇤ ← Revision 1 as of 2025-10-06 19:36:58
Size: 839
Comment: Initial commit
|
← Revision 2 as of 2025-11-03 01:50:41 ⇥
Size: 863
Comment: Link
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 17: | Line 17: |
| Analytic weights (i.e., `[aweight=wt]`) reflect inverse contribution to variance. | Analytic weights (i.e., `[aweight=wt]`) reflect inverse contribution to [[Statistics/Variance|variance]]. |
Stata Weights
Stata supports multiple types of weights.
Contents
Description
Frequency weights (i.e., [fweight=wt]) reflect duplication of rows.
Inverse probability weights (i.e., [pweight=wt]) reflect the probabilities of sample selection across rows. These are appropriate for some types of survey analysis, but complex survey designs generally call for use of svyset.
Analytic weights (i.e., [aweight=wt]) reflect inverse contribution to variance.
'Importance' weights (i.e., [iweight=wt]) reflect relative importance. This is a facility supported for commands to invent uses for, rather than a facility with general use. Any command using importance weights should explicitly define what they mean in that domain.