Skip to main content

New answers tagged

2 votes

Statistical Non-Response and Drop Out

This is an amazingly difficult problem that many have tried to solve. Some of the solutions such as competing risk analysis or counterfactuals are difficult to interpret. Whenever data interruptions ...
Frank Harrell's user avatar
1 vote

Drop-outs in RCT

It sounds then that you can make the missing at random assumption (MAR). That does not mean that it's a good idea to drop subjects with incomplete data (which requires a missing completely at random ...
Frank Harrell's user avatar
3 votes

Multiple Imputation for missing outcomes in Cox regression

What is the question you are trying to answer? E.g. if those that drop-out stop their randomized study drug and you assume that the treatment effect goes away immediately, are you interested in what ...
Björn's user avatar
  • 33k
5 votes

Multiple Imputation for missing outcomes in Cox regression

Where did you see a recommendation to impute outcome data? Imputation is used to create placeholders so that non-missing data are not excluded from analysis. Imputation does not recover information ...
Frank Harrell's user avatar

Top 50 recent answers are included