Controlling the uncontrollable? navigating naturalistic and controlled dosing in cannabis research

Abstract Cannabis research is advancing rapidly, yet fundamental questions persist regarding the optimal design of human studies to empirically assess the potential harms and benefits of cannabis. There exist central methodological differences between studies that rely on controlled dose paradigms, ensuring internal validity, and those that utilize naturalistic, self-administration approaches which better reflect real-world use. Controlled paradigms offer precision and regulatory alignment but fail to capture the complexity of real-world cannabis consumption. Naturalistic designs enhance ecological validity but introduce variability and pose challenges for dosing, safety, and replicability. This perspective article presents two methods: one emphasizing the scientific rigor and utility of controlled dose paradigms for regulatory outcomes (e.g., FDA), and the other advocating for naturalistic models that prioritize real-world relevance of legal market cannabis consumption which can inform policy decisions (e.g., guidelines around high THC concentrate). However, rather than representing competing alternatives, controlled laboratory and naturalistic paradigms occupy distinct positions within the evidentiary framework of cannabinoid research, each addressing scientific questions that the other cannot. An integrated methods section proposes ways to incorporate the best features of both approaches to advance cannabinoid science study design.