Scanning of questionnaires as a tool to identify difficult questions - lessons learned

Background: In 2018, the Luxembourg HBSC team scanned the questionnaires to make the data available faster and to avoid entry errors. Scanning has been shown to be suitable for identifying difficult questions.
Objective: The presentation shows which questions were difficult to answer and what the difficulty was.
Method: The questionnaires were scanned by student assistants and the data was validated by them if the scanning programme did not detect any errors. If errors occurred (e.g. missing answers or multiple answers), then these questionnaires were checked by HBSC team members. This gave us a systematic overview of which questions were difficult to answer. Results 1. The data from 10000 questionnaires were entered in 6 weeks (half the time needed compared to manual entry in 2014). 2. The MVPA question was frequently the subject of multiple answers. This may indicate that these students use the answer scale as a counting aid. 3. Students who state that they have never smoked in their lives often skip the question about tobacco use in the last month. This behaviour can be explained by Grice's conversational maxims. 4. Behaviours indicating that the answers are not serious (crossed-out questions, crosses outside the boxes, fun answers to open questions) are rare.
Conclusions: Scanning is an efficient way to enter many questionnaires in a short time and high quality. Furthermore, it can help to discover difficult questions and to find out what the difficulty is.