Should I use weights when estimating the score?
The recommendation to date seems to be to ignore sampling weights, estimate the propensity score using a logit model (option logit) and match on the odds ratio (option odds)
|
楼主: peyzf
|
1387
1
[回归分析求助] p-score依odds ratio而非predicted probability of being treated |
|
大师 59%
-
|
| ||
|
|
加好友,备注jltj京ICP备16021002号-2 京B2-20170662号
京公网安备 11010802022788号
论坛法律顾问:王进律师
知识产权保护声明
免责及隐私声明


