Imputing outcome data in long vs. wide format

[benji700]

Hi everyone,

I am imputing missing outcome values in a longitudinal dataset using Amelia. I have tried imputing the dataset in long format (with id, time, and value in each row) and in wide format (with id, valuet1,valuet2,valuet3 in each row), and the two seem to produce very different results.

When I have the data in wide format, the imputed data is super consistent with the original data (based on plot() and overimpute() on the Amelia output), but when it's in long format the imputed values don't fit nearly as well and are much more bunched around the mean.

My data is similar in structure to the freetrade dataset used in the intro to Amelia (https://cran.r-project.org/web/packages/Amelia/vignettes/using-amelia.html). I have a dataset of a few hundred people completing a loneliness measure at three timepoints.

Am I justified in just imputing the data in wide format, or am I missing some reason to impute it in a long format?

Thank you for your attention,
Benji