In a project that I am working on, I have reflective and formative indicators. I understand that for the formative constructs I can use PCA (Principle component analysis) for dimension reduction. Can we use PCA for reflective constructs also? If so, I read that certain rotations are only used when the indicators are correlated to each other (Which is the assumption for formative indicators) and certain rotations are used when the indicators are not correlated to each other (Assumption for reflective indicator). So, should we run the analysis separately for formative and reflective indicators?

More Parthesh Shanbhag's questions See All
Similar questions and discussions