Investigating functioning of the #EQ5D -Y-5L in five countries shows that participants can differentiate ordinal response levels, but that measurement quality and scaling functions differ by country and age:
https://www.sciencedirect.com/science/article/pii/S1098301525024830
Since this is conceptually similar to violations of invariance in other measurement models, I wonder whether the #utility #valuation literature needs a more formal framework for this and whether #Psychometrics might help with this?