Most questionnaire surveys to date have ended at the consideration stage, and while there have been some cases in which proposals have been made based on the considerations, the proposals in these cases have remained merely as hypothesis ideas (presenting an idea), and there have been almost no published cases in which hypothesis verification (verification experiments) has been carried out to provide objective evidence for the proposal. In order to clearly show concrete proposals linked to problem solving or task achievement with evidence, experiments that serve as a basis for these proposals are necessary. Conjoint analysis is one option for this. However, most conventional conjoint analysis has been a stand-alone effort, and the analysis is an extremely restrictive (limited) experiment with three characteristics (i.e. all factors are treated qualitatively, b. interactions and quadratic terms are ignored, and c. the experimental design is a factorial or orthogonal design).
This study introduces innovative methods for collecting result data and an experiment (called a virtual experiment in this study) that uses a structural model table and optimal design. It is mathematically clear that in a virtual experiment, it is desirable for the result data to be on a continuous scale. However, suddenly using a continuous scale evaluation places a burden on respondents and produces unreliable data. In response to this, this study proposes an adjusted difference apportionment method.
