Missing observations in longitudinal data

PhanS — Fri, 12 Oct 2018 15:19:17 GMT

Dear All,

I much appreciate your help.

I am analyzing a longitudinal survey data that collected 3 times (T1, T2, and T3). T1=600 observations, T2=500 (15%) was drop-out and T3=400 (13% was drop-out). Thus totally, 28% were a drop-out. Time-unvarying variables like sex, ethnicity did not collect in T2 and T3. An outcome of interests is a continuous variable. The main predictor is a categorical variable.

1) My inclusion criteria, I included only Subjects who had at least 2-time surveys. For instance, to exclude id_102 who missed a variable "ssup_cat" and id_107 who missed all variables. Please kindly see the attached file (I am use a SAS studio).

2) I want to create missing data files (lost to follow-up) in T1 and T2, T2 and T3 and T1 and T3.

How can I tell SAS to do the jobs. Thank you for your time in advance!

Phan S.

Re: Missing observations in longitudinal data

mkeintz — Fri, 19 Oct 2018 00:26:48 GMT

Show examples of the data illustrating all the anticipated conditions, in a DATA step format. The show what you want the result data set to look like.

topic Missing observations in longitudinal data in SAS Procedures

Missing observations in longitudinal data

Re: Missing observations in longitudinal data