High Stakes: The Role of Evaluation in Policy Experiments on Cannabis Regulation

Policy experiments (often referred to as "pilot projects" or "trials") are increasingly promoted as pragmatic governance instruments to generate evidence, manage uncertainty, and support evidence-based policymaking in complex and contested policy domains, where the use and interpretation of evaluation findings carry high political and societal stakes. This development aligns with broader trends in experimentalist governance, which emphasize iterative learning, temporary regulatory arrangements, and evaluation-based adjustment under conditions of uncertainty. At the same time, existing scholarship highlights that the translation of evaluation findings into concrete policy reform is neither automatic nor linear, but highly contingent and shaped by political, institutional, and legitimacy-related conditions. This dissertation examines the role of evaluation in policy experiments, focusing on the conditions under which evaluation can be instrumentally usable for informing policy reform in politically sensitive and complex contexts. The empirical focus is on recreational cannabis regulation, a paradigmatic contested policy field characterized by ideological and normative conflict, multi-level legal constraints, persistent uncertainty, and evolving regulatory and public health challenges related to psychoactive substance use. The study compares two ongoing policy experiments: the Swiss Authorized Pilot Trials with Cannabis for non-medical use and the Dutch Controlled Cannabis Supply Chain Experiment. Building on policy learning, evaluation, and governance literatures, the dissertation develops an analytical framework that distinguishes contextual conditions, evaluative conditions, and legitimacy as a cross-cutting factor shaping evaluation usability. Methodologically, the research adopts a qualitative comparative design combining document analysis with original data collected through a survey and semi-structured interviews involving both experts and stakeholders, in order to examine how evaluation is positioned, perceived, and expected to contribute to future regulatory choices, and to explore the extent to which policy experiments may function not only as learning devices, but also as multifaceted political instruments beyond their explicit learning-oriented rationale.

Policy experiments (often referred to as "pilot projects" or "trials") are increasingly promoted as pragmatic governance instruments to generate evidence, manage uncertainty, and support evidence-based policymaking in complex and contested policy domains, where the use and interpretation of evaluation findings carry high political and societal stakes. This development aligns with broader trends in experimentalist governance, which emphasize iterative learning, temporary regulatory arrangements, and evaluation-based adjustment under conditions of uncertainty. At the same time, existing scholarship highlights that the translation of evaluation findings into concrete policy reform is neither automatic nor linear, but highly contingent and shaped by political, institutional, and legitimacy-related conditions. This dissertation examines the role of evaluation in policy experiments, focusing on the conditions under which evaluation can be instrumentally usable for informing evidence-based reform in politically sensitive and complex contexts. The empirical focus is on recreational cannabis regulation, a paradigmatic contested policy field characterized by ideological and normative conflict, multi-level legal constraints, persistent uncertainty, and evolving regulatory and public health challenges related to psychoactive substance use. The study compares two ongoing policy experiments: the Swiss Authorized Pilot Trials with Cannabis for non-medical use and the Dutch Controlled Cannabis Supply Chain Experiment. Building on policy learning, evaluation, and governance literatures, the dissertation develops an analytical framework that distinguishes contextual conditions, evaluative conditions, and legitimacy as a cross-cutting factor shaping evaluation usability. Methodologically, the research adopts a qualitative comparative design combining document analysis with original data collected through a survey and semi-structured interviews involving both experts and stakeholders, in order to examine how evaluation is positioned, perceived, and expected to contribute to future regulatory choices, and to explore the extent to which policy experiments may function not only as learning devices, but also as multifaceted political instruments beyond their explicit learning-oriented rationale.