I'm writing up a case study and would like to calculate the consistency of a patient's responses to the same set of 'yes/no' questions asked at two different time points.
I had hoped to use Cohen's Kappa as a measure of agreement, however I'm not clear whether this is valid- most of the examples I've found have had two raters assessing an invariant stimulus (e.g. video or audio recordings of an interview), though I've found some research where kappa is calculated when the same patient has de novo assessments by two raters.
However, I have repeated measurement of a patient with one rater. Is it appropriate to use kappa in this situation?
Andrew Wood From the sounds of it, your study doesn't have any raters. The data is coming from the patients, not the researcher - so you wouldn't use Cohen's kappa. Cohen's kappa is used when you have two independent raters.
I suggest using Cronbach's alpha instead (assuming your items are all measuring the same underlying variable). See the below thread for further details: https://www.researchgate.net/post/Is_it_correct_to_use_Cronbach_s_Alpha_to_assess_the_internal_consistency_of_a_questionnaire_with_binary_data_yes_no
- Badges
- Science topic
- Similar topics
- Philosophy
- Ethics