Functional unblinding happens when participants or staff can work out who received the real drug from its effects, even though the trial was designed so that no one would know. Nobody opens the code; the experience gives it away.
Psychedelic trials show it often. In a 2022 alcohol use disorder trial, about 94% of participants guessed their assignment correctly (Bogenschutz et al., 2022). In the 2026 EPISODE depression trial, 86% correctly identified the higher dose (Mertens et al., 2026). At an FDA advisory committee in June 2024 on MDMA-assisted therapy, members who voted against it pointed to functional unblinding (FDA minutes, 2024).
Why it matters: once people know their group, expectancy can push scores up in the drug group and down in the control group. A 2021 review argued that effect sizes are therefore likely overestimated (Muthukumaraswamy et al., 2021). Fixes under study include active placebos, low-dose comparators, raters who never meet participants, and asking everyone to guess their group so the problem can at least be measured.
Sources
- Bogenschutz et al., JAMA Psychiatry, 2022
- Mertens et al., JAMA Psychiatry, 2026
- FDA advisory committee minutes, 4 June 2024
- Muthukumaraswamy et al., 2021
Last checked .