Hi,
To understand participants' life satisfaction, 5 questions have been asked. In each question, there are 5 options, from disagree to agree. Please suggest me how to create a single variable from these 5 categorical (ordinal) variables through MCA preferably in STATA or else in SPSS.
Dear Shishirendu Ghosal ,
I have performed MCA using Stata and it is pretty easy.
(1) You conduct your MCA as
mca varlist
(2) You predict the first score as
predict score1
(3) Remember that you can change the sign for convinience of interpretation (it is just a linear combination). For example,
replace score1 = -score1
I hope it helps.
You can mark the question as solved if it is the case.
Best,
José-Ignacio
The fascinating question is why you are using MCA.
I would start by verifying that the five questions meet a definition of a unidimensional scale. In Stata, have a look at Article Nonparametric Item Response Theory Using Stata
In SPSS, have a look at Stata.
To reiterate a point frequently made on ResearchGate, Cronbach's alpha does not measure unidimensionality, though it assumes it.
Were the responses unidirectional? I would suggest checking internal consistency first, before proceeding with MCA.
Ronán Michael Conroy sir,
Thank you for sharing the article. Will go through it.
- Badges
- Science topic