S1E16: Missing Data: (IF EPISODE=16 THEN EPISODE=-999)
Quantitude
00:00
The Importance of External Validity in Modeling
In the 1980s when we didn't have multiple imputation, there were a number of ways to look at data. We could use list wise deletion - if anybody was missing something on anything, you're out. Problems with that tough to invert; they're often not positive definite. Some clever ones we can use a mean imputation. And then the biggest one is what is the poke in the eye? It really does relate to internal and external validity.
Transcript
Play full episode
