I want to carry out a survey on medical students with self administered questionnaire (21 statements with 3 sub groups with YES, NO, Don't Know response). I would like to score 2 for correct response, 1 for Don't Know response and 0 for wrong response. I would like to test the Reliability of the questionnaire. For that, i would like to do a pilot study. My question is how many students should be included in pilot study to test Cronbach's Alpha?
I have gone through some article including this https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4205511/ but could not find my answer.
Dear Himel,
Cronbach's Alpha does not depend on sample size. If you take e.g. 16 for you pilot it is enough but yes and no is a bad scale. make at a 5-point scale but preferable a 7-point scale.
I am suggestion; pilot test sample size may be 30. Test simple size is 21 item x5 (min)=105.
Monte Carlo is a non-parametric measure, if you want your pilot study to be normally distributed then you can use n= 30. This is not necessary.
https://en.wikipedia.org/wiki/Monte_Carlo_method
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire