The DDME project aims at a novel understanding of future democracy. We first conducted a large survey in Germany (n= 2632) and the US (n=2637) to capture citizens´ understandings of democracy (based on the open-ended question at the beginning “People have different ideas about democracy. In your own words, what does democracy mean to you?” as well as on closed-ended questions). Regarding citizens´ understandings of democracy, we employed Biterm Topic Modelling (a technique using unsupervised machine learning which is particularly apt for the analysis of short texts) in combination with sentiment analysis and in-depth human coding. We found elements of “textbook democracy”: both German and American citizens emphasized traditional principles such as “rule of the people” and “freedom” (as emphasized by Dahl and Schumpeter). Overall, while citizens´ understandings of democracy are not “random”, they still leave ample room for creative input what democracy is and could be.
In a further step, we focused on concrete governance tools. We identified four basic schemes: a participatory scheme (involving lottocratic and/or direct democratic devices), a representative, an expertocratic and an executive scheme. A key component of the theoretical and empirical analysis is that citizens might prefer combined or “blended” models of governance - mixing representative with participatory, expertocratic and executive elements - to singular schemes (such as pure representative or pure direct democracy). We developed and tested novel survey techniques to capture such complex governance preferences, namely constant sum and conjoint analysis. Constant-sum analysis allows participants to allocating 100% of decision-making powers in increments of 5% to the four governance models (representative. participatory, expertocratic and executive). This approach captures preferences for mixed or blended governance models more effectively than traditional techniques (such as latent class analysis). Our new and modular conjoint design enables respondents to directly compare different decision-making models, including “blended” forms of governance. In the pre-test of the large engagement surveys (randomly assigning participants to a Mars and Earth setting), we presented respondents with four main decision-making institutions: a parliament (representative mode), a randomly selected citizen forum (participatory mode), an expertocratic committee (technocratic mode) and a democratically elected assertive leader (executive mode). We further innovated by combining the four main institutions with additional qualifications and institutional options, such as special consideration of expert or public opinion (including input by a minipublic), final decisions made by referendum or by the main institution and a final check by constitutional court.
A key goal of DDME is to put citizens into hypothetical and counterfactual scenarios to explore what their “true” governance preferences are (absent any constraints). We use “Mars” as a metaphor and a heuristic aid to stimulate imaginative and counterfactual thinking about institutional architectures. For the main experiments, we conceptualize “Mars” as an institutional “blank slate” as well as a mass society where the basic infrastructure is set (except the political infrastructure). We have run an extensive cognitive pre-test with ten citizens in Germany and eight citizens in the US (with different levels of education, age and gender) to validate what they understand by our definitions of “Mars” and “Earth”; we find that the treatments are correctly understood. We have (pre-)tested our novel engagement survey with adolescents, university students and with large samples in Germany (n=974) and the US (n=1012). Results show that in the “Earth” condition, citizens (including adolescents) stick to the representative system as the main institution, even though there is some desire for combining representative institutions with referendums as well as for checks-and-balance institutions (Schwaiger and Bächtiger 2024). By contrast, in the “Mars” setting respondents (especially in Germany) have a tendency to prefer participatory governance schemes as the main institution. At closer inspection, however, what citizens actually want on “Mars” is - as in Ancient Athents - that different goverance institutions are jointly involved in decision-making. A large engagement survey with 5000 participants per country will be conducted in Germany and the US in fall 2024 and in in India in spring 2025.
Finally, we have also invested considerable time to design the deliberative experiments. We determined (based on recent experiences) that it is impossible to combine a large representative sample (with sufficient statistical power for proper subgroup analysis) with a deliberative and co-creative design. Too few people will be willing to fill out both extensive questionnaires and engage in co-creative designing. Therefore, we will conduct two types of experiments: (1) minimal engagement surveys with large samples (about 5000) where participants engage with pro-and con-arguments and make individual design choices; and (2) deep engagement designs where smaller numbers of (maximally diverse) participants (about 150) deliberate together based on the input of the minimal engagement survey and produce more concrete and more detailed proposals how different governance schemes can be optimally weaved together for institutional renewal. The deliberative experiments will be conducted in early summer 2025.
