Hi I would be grateful for an opinion on the following
When participants just complete early demographic questions in a survey and don't progress to the actual questions, is this classified as NMAR data? Given that I still have sufficient number of completed surveys if the incomplete ones are excluded to meet the requirements of power calculations, can I just exclude them. Are there any valuable references that will show this is acceptable?
Thank you in advance
Colin
Here is a link to assist you in identifying MNAR. I'm not sure that you can determine this, though, if they only completed the demographics.
https://www.ncbi.nlm.nih.gov/books/NBK493614/
You could also consider listing Nonresponse bias/Item Nonresponse Error, but those typically refer to random missing data. If they did not continue after demographics, we typically exclude them and just state 'x number of participants were excluded for not completing the majority of survey content' or 'x potential participants began the survey but were excluded for only completing a small portion,' etc. You'll want to identify how many began vs. how many total you ended with. Still, it's worth noting that they never completed more than the demographics, so the reader can understand how your decision was the most appropriate option.
Thank you Deborah, your answer is v helpful. I will access your link.
Thank you Deborah, your answer is v helpful. I will access your link.
- Badges
- Science topic
- Similar topics
- Statistics
- Survey