Getting a dataset with imputed values for multi Item questionnaires using MICE

Question

I have multiple multi item questionnaires (surveys)on a likert scale with missing data, My goal is not to do further analysis but to generate an imputed dataset for my collaborators to use. I am trying to understand how to handle this?

I have found a great workflow to accomplish imputing data on multi item questionnaires using passive imputation: exercise: 8.2.3 Passive multiple imputation in R, https://bookdown.org/mwheymans/bookmi/missing-data-in-questionnaires.html#passive-multiple-imputation-in-r

From what I understand it is discouraged to take an average of multiple imputed datasets generated from MICE (violates Rubins rule) and is recommended use the with() and pool () function to do further analysis.

As mentioned earlier, my goal is not to do further analysis but to generate an imputed dataset. I see that there is an option to get a "long" dataset with all the imputed values for a subject or choose any one of the datasets from multiple imputed datasets.

For anyone who has worked with these kind of data before, I am looking for recommendations on what to do next? get a long dataset or choose any dataset? I am open to other workflows or packages that impute missing data on multi item questionnaires.

Answer 1

For anyone else looking for an answer, long data is recommended