I need to analyze a survey about entrepreneurship which has around 50 categorical variables.

Therefore, after some univariate analysis, I want to apply Multiple Correspondence Analysis (MCA) in order to reduce dimension. So far, I have seen the videos about MCA from https://www.youtube.com/playlist?list=PLnZgp6epRBbTsZEFXi_p6W48HhNyqwxIu and read some sections from Exploratory Multivariate Analysis by Example Using R book.

My questions regarding MCA are:

  • Is it necessary to fully understand CA first?

  • How do I work with multiple choice questions? (Should I create binary variables? Or Should I count how many categories each individual marked?)

  • For overlapping questions (To answer question "B", the respondent had to answer "Yes" for question "A". Those who have answered "No" to question "A", should I put "No" in "B"?)

  • Is it a good idea to make questionnaires with multiple choice or overlapping questions? (It seems that it is not, because book examples don't bring this kind of questions).