Keep in mind that I'll be working from R, if that's important, but let me set up the problem I'm hoping to solve. First, I have 15 4x4 pairwise comparison matrices per person. They have an 8-point scale between the objects being compared. I'm curious if a) even a single matrix violates AHP assumptions, and b) would averaging 15 such indices result in a meaningful total index?

I am not interested in AHP's formal intended purpose. This design is not built for AHP, but I want to capture consistency or inconsistency, and it appears that AHP is the primary perspective with a formalized set of consistency indices.

More Christopher Gallagher's questions See All
Similar questions and discussions