Attrition is the worst, especially in longitudinal panel datasets. One trick I’ve used is to compare attritors vs. non-attritors early on—if they’re similar, you might be okay.
For gaps, `zoo` in R has some handy functions for time-series interpolation. And if you’re drowning in data, consider sampling or using sparse matrices to cut down the size.
Also, +1 for `data.table`. It’s a game-changer for speed.
For gaps, `zoo` in R has some handy functions for time-series interpolation. And if you’re drowning in data, consider sampling or using sparse matrices to cut down the size.
Also, +1 for `data.table`. It’s a game-changer for speed.