Hi,
I'm analyzing survey questions (25 items) scored each 1-7 on likert scale. How to deal with respones that are missing from some of the items (i.e. some individuals didn't answer all 25 questions and left few blank). My design is pre-test post-test with paired sample (the sample is not random). Do I calculate the total scores for those individuals but pro rate them based only on the number of items answered. so if they answered 21 questions, the scores will be total/21?
I don't believe deleting them is a good idea. Does it worth imputing the scores, I'm not very familiar on how to impute exactly but I have a rough idea.
Thank you, looking forward for your help.
I suggest not to consider incomplete questionnaires for the purposes of analysis.
- Badges
- Science topic
- Similar topics
- Statistics
- Survey