Recent research challenges traditional singular models of government (e.g., representative vs. direct democracy) and advocates for mixed models integrating representative, participatory, expertocratic, and executive elements (e.g., Pilet et al. (2023), Schwaiger & Bächtiger (2024)).
This study advances the measurement of preferences for mixed governance systems by using a constant sum question to directly capture trade-offs and nuanced preferences across multiple governance models. Analysing two representative samples of 5,000 respondents from Germany and the US, the results reveal a strong preference for mixed models, with 93% of respondents dividing decision-making power among at least two models, and most incorporating all four (representative, participatory, expertocratic, and assertive leader model).
Further analyses confirm the validity of the constant sum approach, and show its alignment with other preference measures, such as Likert-scale items, a single-choice question, and results from conjoint analyses. Participants' justifications for their model combinations indicate considered decision preferences rather than simplistic gut reactions or non-attitudinal responses. The constant sum method also differs from the Likert scale items in terms of its predictive power and response bias, such as acquiescence. In addition to directly capturing preferences for specific governance model combinations, the nuanced structure and determinants of these mixed preferences become even more apparent when combined with fuzzy clustering and regression analysis. This paper demonstrates the value of constant sum questions in advancing the measurement of preferences for mixed governance systems and their underlying trade-offs.