I has questionnaire with likert scale, The likert scale goes from: 1=never, 2=seldom, 3=sometimes, 4=often, 5=very often and 6= not applicable.
I want to know how many time the student engage in behavior..
Can i convert likert scale to dichotomous, consider to :
never and not applicable is 0 (not engage ),
seldom, sometimes, often and very often is 1 (engage)?
Is there any studies regarding to this thing ?
What are the limitations if you convert likert scaled data to dichotomous ?