Psychedelic Phase 3 readouts are reported the way most drug trials are reported: a headline effect size, a p-value, a comparison to placebo. For this class of therapy, that framing is close to useless on its own. The structural features that make psychedelics promising — a powerful subjective experience, motivated participants, intensive therapeutic support — are the same features that make a naive placebo comparison hard to trust. This primer is a reference for reading these trials on their own terms: what each design choice is trying to control for, what a given result can and cannot tell you, and the specific questions to ask before you believe a headline number.
The blinding question
Functional unblinding is near-total in these trials: most participants correctly guess their arm. [Stub: explain why blinding fails structurally, what an unblinding analysis measures, and why a failed blind is expected rather than disqualifying.] The question is never whether the blind held — it didn’t — but what the sponsor did to measure and account for it.
Expectancy and how to control for it
Expectancy is the single largest confound in the field. [Stub: define expectancy effect, describe how enrollment and preparation amplify it, and survey the control strategies a serious protocol uses.] A trial that ignores expectancy is not measuring drug effect; it is measuring hope plus drug effect, with no way to separate them.
Dose-response designs
A dose-response signal is one of the cleanest ways to demonstrate a pharmacological effect when blinding is impossible. [Stub: explain how comparing low/medium/high doses isolates drug effect from expectancy, and note the designs that have used this well.]
Active comparators
A psychoactive comparator can hold expectancy roughly constant across arms. [Stub: discuss niacin, low-dose active, and other comparators; what each preserves and what each gives away.]
Mechanistic biomarkers
Objective markers can corroborate a subjective readout. [Stub: cover neuroimaging connectivity measures, neuroplasticity markers, and physiological correlates — what they add and their current limits.]
Attrition handling
Differential dropout can manufacture or destroy an effect. [Stub: explain intention-to-treat vs per-protocol, how missing-data imputation choices move the result, and what to look for in the dropout table.]
Reading the secondary endpoints
Secondaries reveal durability and breadth that the primary often hides. [Stub: discuss durability of response, functional outcomes, and the difference between a statistically and clinically meaningful secondary.]
The questions to ask
A short checklist a reader can apply to any readout. [Stub: was unblinding measured; was expectancy controlled by design; is there a dose-response or biomarker corroboration; how was attrition handled; do the secondaries support the primary.]
Behavioral Wire’s research desk publishes methods references and data analyses. Reporting is held to standards of attribution and verification described in our methodology.