I wonder if anyone has any insights into this issue?
The questionnaire I am using has come back with about 16% missing data
overall, and the missingness appears to have a pattern.
The questions come in sets of three (e.g. 2a, 2b, & 2c) and sometimes
participants answer the first question in the set and then skip the other
two (for all items roughly 5% skip the first item in the set, and about 12%
skip the next two). My guess is that they are skipping for one of two
reasons: 1/ they are skipping items instead of putting a score of zero
(0='no problem'), or 2/ they think the question is multiple choice (i.e.
they think they are supposed to choose one of a, b, or c) though the
questionnaire states clearly they should answer all questions, and an "N/A"
option is included.
I would like to hear opinions my options e.g. can I use the dataset, but
leaving out anyone who has missed data? Or do I need to scrap the whole
dataset? Or ask the participants who skipped to say why they did so?
Thanks for any feedback...
|