How can I categorize the Adherence in Chronic Disease Scale data into ‘poor’ and ‘good’ adherence categories for the purpose of multivariate logistic regression analysis? the nature of ACDS questionnaire are low(0-20), moderate(21-26) and high adherence(>26) but i want to categories into two non adherence and adherence and i trying to categories >26 score is dherence and less than or equal to 26 non adherence but dherence much more low compared to non adherence
What's wrong with sticking to the original low, medium and high categorization? Inclusion of the middle category could be helpful in identifying a "good enough" level of adherence? Given a large enough sample you could experiment with multiple levels in a search for "good enough".
You should look into ordinal logistic regression (or sometimes called ordered logit models).
- Badges
- Science topic
- Similar topics
- Psychology
- Mental Processes
- Learning