AMOS handling missing data from follow up study (where survey not completed)

Gia2

New Member
#1
I am conducting a study which requires participants to complete two questionnaires a month apart. A total of 300 participants completed the initial questionnaire, but only 170 completed the second.

I'm not sure how best to handle this in AMOS. With there being so much missing data I don't want the values to be estimated. I believe that with pairwise deletion, all cases are included in the analysis, but when there is missing data for a variable it excludes the case for that variable only.

Any ideas as to what method would be best in this situation?
 

Lazar

Phineas Packard
#2
Does AMOS use full information maximum liklihood [FIML] by default? If so don't do ANY deletion just run your model and FIML will handle missing data for you.
 
#3
I am conducting a study which requires participants to complete two questionnaires a month apart. A total of 300 participants completed the initial questionnaire, but only 170 completed the second.

I'm not sure how best to handle this in AMOS. With there being so much missing data I don't want the values to be estimated. I believe that with pairwise deletion, all cases are included in the analysis, but when there is missing data for a variable it excludes the case for that variable only.

Any ideas as to what method would be best in this situation?
An important question is, what are you trying to do with survey data? i.e. what questions you asking, and what analysis? This can help determine how you deal with missing data. I'd be wary of estimating missing-values given the amount of missing data, especially if you think there could be a reason for the missing data.