Several researchers I heard interviewed about psychedelic medicine mentioned that they use the substances themselves. That set me wondering: can someone remain an unbiased observer of something they have personally embraced? It turns out science has names for this problem - and tools for it. A quick survey, using psychedelic research as the worked example.
I have been thinking about the biases that can creep into scientific research - even research done by well qualified people at leading institutions. What set me on this path was a run of podcasts featuring researchers studying psychedelic medicine. The narrative they told was consistent: these compounds were researched in the 1960s; that research was shut down for political reasons and became very difficult to conduct; and they now show promise as treatments for a range of mental health conditions.
Let me be clear at the outset. I support research into treatments for mental health, and if these compounds prove genuinely helpful, I am in full support of them as medical interventions, administered under proper supervision. I have no first-hand knowledge of the historical politics, so I make no comment on that part of the story.
What caught my attention was something else. More than one of these experts mentioned that they use psychedelics themselves. And that raised a question I couldn't shake: are they still unbiased observers? Can we trust conclusions drawn from their research? I want to be precise about what I am and am not asking. I am not questioning these researchers' qualifications, and I am not casting doubt on their motives - they sounded sincere to me, and well equipped for their jobs. The question is about the scientific method itself: what happens to objectivity when the investigator is personally invested in the thing being investigated?
I couldn't fully articulate the concern to myself, so I went digging. Is this a known problem? Does it have a name? And was my impression even representative, given I'd heard only a handful of interviews?
The problem has names - and they are many
It turns out the phenomenon I was circling is well charted. A quick tour:
Experimenter expectancy. A researcher's beliefs can leak into a study without any dishonesty at all - through how questions are framed, how participants are treated, how ambiguous data gets interpreted. Psychology has studied this since the 1960s; the APA defines it as the experimenter's hypothesis unintentionally influencing the outcome.
Demand characteristics. The mirror image on the participant's side: people in studies pick up cues about what the researcher hopes to find, and - often without meaning to - deliver it. This matters enormously wherever outcomes are measured by asking people how they feel.
The allegiance effect. In psychotherapy research it is well documented that a researcher's loyalty to a treatment predicts how well that treatment performs in their own trials - a meta-analysis of the field found a moderate, robust association between researcher allegiance and outcomes. Nobody is accusing anyone of fraud; the bias operates through a hundred small choices.
"Going native." Anthropologists coined this one for the field researcher who becomes so immersed in the community being studied that critical distance disappears. The parallel to a scientist who has personally adopted the practice under study draws itself.
Epistemic contamination - and its opposite. Some argue that a profound experience can rewrite a person's frame of reference so thoroughly that they can no longer evaluate it neutrally; they now belong to it. Others argue the reverse: that a researcher who has never had the experience is like a colourblind scientist studying colour perception, missing something essential about the very phenomenon in question. Both arguments sound valid to me and seem to apply to this specific case.
Why it matters here
Psychedelic research is unusually exposed to all of the above, for one structural reason: the key outcomes are subjective. Depression scores, well-being, "mystical experience" questionnaires - all self-reported, which is exactly where expectancy and demand characteristics do their work.
And the field's standard defence - the double-blind, placebo-controlled trial - struggles here. A 2021 review of blinding in psychedelic trials found that participants and staff can usually tell who received the active drug; you generally know whether you have had a psychedelic. Researchers call this functional unblinding. When the U.S. FDA's advisory committee reviewed MDMA-assisted therapy for PTSD in June 2024, members voted 10–1 that effectiveness had not been demonstrated, citing functional unblinding and expectancy effects among, the great majority of participants correctly guessed which arm they were in. The FDA declined approval that August and asked for further study. Whatever the ultimate merits of the treatment, the regulator's concern was the same one I had stumbled into from my armchair.
Was my impression correct?
Partially - and this is the encouraging part. My sample of podcast guests was not representative. The field actually spans a spectrum. At one end are open advocates: Rick Doblin, founder of MAPS, has never presented himself as a neutral observer - he is explicit that he is an advocate. The late Roland Griffiths of Johns Hopkins, who did as much as anyone to restart this research, came to it as a sceptical behavioural pharmacologist and only late in life spoke publicly about his own experiences. At the other end are strict pharmacologists like David Nichols, who spent four decades on the chemistry and receptor science of these compounds while keeping the work squarely objective.
Warnings even come from inside the field. Matthew Johnson, a psychedelic medicine researcher, has published explicitly on the pitfalls of psychedelic research - cautioning colleagues against importing their personal spiritual beliefs into trials and against researchers drifting from scientist toward guru. So I rest assured that the researchers are not uniformly personal users, and the community is actively arguing about exactly the bias that worried me. A field that polices itself in print is a field taking the method seriously.
How science defends itself
None of these biases is unique to psychedelics, and none is fatal. There are options, and reviewers in this field have suggestion how to fix it: active placebos that mimic some drug effects; measuring - rather than assuming - how well blinding held and what participants expected; independent raters who never meet the participants and don't know their assignments; objective endpoints such as brain imaging and biomarkers alongside the questionnaires; preregistered designs that stop the goalposts from moving; and honest disclosure of the researcher's own position, so readers can weigh it. There is even an ongoing debate about whether researchers should disclose personal use in their papers - transparency versus the risk of prejudicing readers - which is a good sign that the problem is being addressed.
In Conclusion
I began with a suspicion I could not name and ended somewhere more hopeful. The bias I sensed is real, it has names, the regulator has acted on it, and the field's own members are its loudest critics. That is not a story of science failing. It is a story of science doing the thing it was designed to do: refusing to let the sincerity of the investigator substitute for the rigour of the investigation.
That is the general lesson worth carrying out of this particular field. Every researcher is a human being with experiences, loyalties, and hopes; objectivity is not a personality trait some people have. It is a discipline, built out of method - blinding, controls, preregistration, adversarial review - precisely because we know we cannot simply trust ourselves. Trust in science, properly placed, is trust in that machinery, not in any individual's freedom from bias.
And if research conducted with that machinery in good working order shows that these compounds genuinely help people suffering from mental illness, then we should welcome the result and build those treatments - for use under the supervision of competent medical professionals, where they belong.
References
- APA Dictionary of Psychology, "experimenter expectancy effect" - apa.org
- APA Dictionary of Psychology, "demand characteristics" - apa.org
- Munder et al., "Researcher allegiance in psychotherapy outcome research: An overview of reviews" - Clinical Psychology Review (2013)
- Muthukumaraswamy, Forsyth & Lumley, "Blinding and expectancy confounds in psychedelic randomized controlled trials" - Expert Review of Clinical Pharmacology (2021)
- "In a setback for psychedelic therapy, FDA advisers vote against medical use of ecstasy" - Science (2024)
- "FDA rejects MDMA-assisted therapy for PTSD treatment" - The Conversation (2024)
- Johnson, "Consciousness, Religion, and Gurus: Pitfalls of Psychedelic Medicine" - ACS Pharmacology & Translational Science (2021)
- Aday et al., "Great Expectations: recommendations for improving the methodological rigor of psychedelic clinical trials" - Psychopharmacology (2022)
- "A Psychedelics Pioneer Takes the Ultimate Trip" (Roland Griffiths interview) - The New York Times (2023)
- Rick Doblin - MAPS · TED (2019)
- Nichols, "Psychedelics" - Pharmacological Reviews (2016)
These thoughts are my own. I had help with grammar, structure, and research from Claude.