Who Guessed the Assignment? Expectancy as a Measurement Problem

A trial can be described as double-blind even when many participants suspect which treatment they received. Psilocybin makes this problem unusually visible because the acute experience can differ substantially from a control condition. The important question is not whether a study used the word “blind,” but what information participants, facilitators, and outcome assessors could infer, and when they could infer it.

Expectancy is not a synonym for an imaginary improvement. Expectations can influence how symptoms are experienced, described, and rated. They can also influence engagement with support sessions and willingness to remain in a study. Understanding those pathways helps readers interpret the site's depression evidence overview without assuming either that all benefit is pharmacological or that all benefit is placebo.

Allocation concealment and masking do different jobs

Allocation concealment prevents recruiters from knowing the next assignment before a participant enters the trial. Masking attempts to limit knowledge of the assignment afterward. A study can succeed at the first and struggle with the second. Proper randomization still matters when masking is imperfect, because it helps balance starting characteristics. But it does not prevent later expectations from diverging between groups.

In the 2023 Raison trial, psilocybin was compared with niacin alongside psychological support. Niacin was intended as an active control, but an active control does not automatically reproduce a psychedelic experience. The report identifies functional unblinding as an interpretive issue. Readers should distinguish the design intention to mask treatment from evidence that masking actually remained effective. [1]

Measure expectations before the experience

Before treatment, participants may have different hopes, fears, and beliefs about psilocybin. Prior psychedelic exposure, media coverage, and preference for a particular assignment could affect those expectations. Measuring them before dosing is useful because a later rating can already reflect perceived treatment effects. A baseline expectation measure does not solve the problem, but it makes an otherwise hidden influence more visible.

If expectations are collected, ask whether they were similar across groups and whether the analysis was planned in advance. An association between optimism and improvement does not prove that optimism caused the improvement. More severe symptoms, therapeutic engagement, or other characteristics may influence both. Likewise, an absence of a statistically significant association in a small sample does not establish that expectancy played no role.

Timing changes what a treatment guess means

A guess made during an acute session may reveal sensory clues. A guess made weeks later may reflect symptom improvement, disappointment, conversations with staff, or all of these. Researchers therefore need to describe when guesses were collected and whose guesses were recorded. A single end-of-study question cannot cleanly separate acute recognition from later inference based on outcomes.

Correct guesses are also not a simple numerical estimate of bias. Someone might correctly identify treatment because an effective drug produced recognizable effects. Another person might report improvement because they believe they received the desired treatment. Both processes can occur together. A masking assessment provides useful diagnostic information, not a formula that can simply be subtracted from the observed treatment effect.

Independent raters help, with limits

Separating outcome assessors from session facilitators can reduce information leakage. Standardized interviews and instructions not to reveal treatment experiences can also help. Yet participants may inadvertently disclose a vivid session or disappointment about assignment. A blinded rater can still hear a symptom report shaped by participant expectations. Independent assessment reduces one pathway of bias rather than eliminating every pathway.

The distinction matters when comparing clinician-rated and self-reported outcomes. Agreement between them can strengthen confidence that improvement was not confined to one instrument, but both can be influenced by overlapping information. Disagreement may reflect measurement differences rather than dishonesty. The useful response is to examine the assessment procedures and timing, not to declare one kind of rating automatically objective.

Comparator choice creates trade-offs

The Goodwin trial used different psilocybin dose groups, including a very low-dose control. Such a design can address dose-related questions, but it does not guarantee that participants experience the groups similarly. A lower dose may have some effects of its own, while a higher dose may be easier to recognize. The study's results therefore need interpretation within that specific comparison. [2]

An established antidepressant comparator poses a different challenge. In the 2021 psilocybin–escitalopram study, both groups received psychological support and different medication regimens. That design made a clinically relevant comparison, but distinctive experiences and preferences still complicate interpretation. No single control condition perfectly matches acute sensations, treatment expectations, and every potentially therapeutic component. [3]

What a stronger report should show

Look for baseline expectations, treatment guesses collected at meaningful times, assessor separation, procedures limiting accidental disclosure, and balanced descriptions of both interventions during consent. Also look for acknowledgement that disappointment may affect retention or reporting. The site's discussion of set and setting provides context for why interpersonal cues deserve attention alongside the capsule.

The fairest conclusion is rarely “blinding failed, so nothing was learned.” Nor is it “randomized, therefore expectations cannot matter.” Trials can provide valuable evidence while leaving uncertainty about how much observed benefit reflects pharmacology, expectations, support, or their interaction. Naming that uncertainty is a way to improve the research question, not a reason to stop asking it.

Disclosure: It provides research education, not medical advice or a recommendation to use psilocybin.

Editorial disclosure: Generated pictures are illustrative, not study participants. Consult a qualified professional for personal care.