0
$\begingroup$

I'm doing a research where I applied a questionnaire with 67 items measuring 5 variables.

The issue is some questions are in Likert scale (0 to 10 and 1 to 5), some Yes, No, some open questions, some ordinal, some nominal. How can I mix those types into a single one in order to have each of those 5 variables?

Also the questionnaire also contains conditional items. E.g., if someone chooses "No" for question 15, he or she will skip question 16 and jump to question 17.

For example:

1- Level of education (0-Nothing, 1-Elementary, 2-Secondary school, 3-High school or equivalent 4-University 5-Master, PHD

12- Type of cell (1-Shared cell, 2-Dormitories)

15-How many people are imprisoned with you in your cell? (Number... from 0 to 100)

24-Do you take part of social or recreational programs/activities? (1-Yes, 2-No)

35-From 0 to 10, what your level of stress (Likert with 11 elements from 0 to >10)

46-Do you think rules are fair for you and fellow inmates (Likert with 5 elements, from strongly agree to strongly disagree)

57-Do you get psychologic exams? 1-Yes, 2-No If not, skip question 58, go to 59...

66-In your opinion, what's the main reason behind inmate misbehavior? ___ open answer

This is a small summary, i have 67 questions of those types...

Because now I think I cant do the Cronbach alpha. Most of the examples i've seen contain like 4 Likert questions, 40 subjects, in my case my questions are different, not in this same scale.

I'm using Stata.

$\endgroup$
  • $\begingroup$ I suppose the questionnaire has scoring rules, that tells you what to count and how to obtain the 5 scales you are measuring $\endgroup$ – rep_ho Oct 12 '17 at 5:44
  • $\begingroup$ yes, but each variable has different types of items. That's my concern. For example, for variable Communication, there are some Likert, some yes/no, etc $\endgroup$ – Colonel G Oct 12 '17 at 5:56
0
$\begingroup$

In Stata, my view is that you can use alpha q1 q2 q3 q4, std

That std will help you no matter the type of item.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.