Hierarchical Information Integration Experiments and Integrated Choice Experiments

Abstract

When conjoint experiments are applied to study complex decision making that involve many attributes, this often results in problems of information overload and respondent burden, potentially jeopardizing the validity of such experiments. To avoid or reduce the impact of these potential problems, Hierarchical Information Integration has been suggested. The key notion is to classify the large number of potentially influential attributes into a smaller set of decision constructs, construct separate experimental designs for each of these constructs and in addition a bridging design that allows the scaling of all part‐worth utilities into a concatenated utility expression. The basic approach suggested for preference measurements has been elaborated for other measurement tasks and the original design strategy has been refined into an alternative approach. This paper summarizes these developments and briefly discusses aspects of respondent burden and validity.

Notes