I realised that a block of participants are all missing data on one question on my Friendship Questionnaire. Where this happened on the AQ I filled in the missing data with the mean scores, however this one is ranked data, i.e. the participants were asked to rank order three options. I’ve looked at filling in the score with the mean but it doesn’t make sense because of how the survey is scored (5 points if they rank a certain option first, 0 if they rank either of the other options first). Is there any way I can fill in this data?
Dear Hannah
Since it is related sample data, if you contact the respondents it would be better. Otherwise, and as the tests should be related statistical tests.Then, the incomplete surveys should be discarded /neglected, and the remaining (>30) surveys are sufficient for statistical analysis.
Regards,
Zuhair
Hi Hannah,
Missing data is always a big problem, as it is "non-responsive".
Subposing by the mean is a procedure that some use to save data unanswered, but it is problematic and can "reveal a world that does not exist."
We analyze the data to seek, almost always, differences. The use of averages in the place of the missing minimizes these differences.
The best procedure is to delete the subjects with missing data and collect new answers.
Hope this help you
Best Regars
Dirceu
- Badges
- Science topic