I want to create a survey to assess relevance of health technology. Basically the survey will assess relevance by identifying whether the technology is a good fit for an individual.
I am thinking about using conditional branching.
For example, if a user suffers from stress, the items will ask a user about the extent to which a technology meets user's stress-related needs (e.g., providing stress techniques, etc.)
If a user does not suffer from stress, these stress-related questions will be omitted.
There will be multiple sections like that (e.g., physical pain, stress, interference with daily life, etc.) based on user's experiences with illness.
How could I assess validity and reliability of the survey instrument with conditional branching?
Initially, I was thinking about using a Likert scale for all items (1-disagree to 5-agree) and adding an NA option. Then, I wanted to use IRT to treat NA as a missing value. And then, I could run factor analysis (EFA and CFA) to validate.
What is your opinion on this? Please let me know..