The Psychedelic Five

Research report · September 2026

Audience segmentation · US adults

The Psychedelic Five

Americans don’t split into two camps on psychedelics. They split into five audiences, three of which sit in a middle that a national average makes invisible. Here is who they are, and how to reach each one.

1,527 US adults surveyedFielded November 20239 questions, 5 audiences

Read the findings Which one are you? Take the quiz

Out of every 100 American adults

Scroll ↓

01 · At a glance

Six things to know

Five audiences, and three of them are in the middle.

Advocates (17%) and Opponents (36%) hold settled views at opposite ends. Between them sit three groups a poll average blurs into one: Conflicted Insiders (12%), who know psychedelics well and are unconvinced; Curious Moderates (22%), who know little and already lean favourable; and the Disengaged (12%), who have not engaged at all.

22%

The biggest movable audience is warm already.

Curious Moderates are the largest audience that has not settled. Only 5% call themselves very or extremely familiar and 79% describe their feelings as neutral, yet 54% judge the benefits of psychedelics to outweigh the harms once asked about specifics.

4.4

Absence is not opposition.

The Disengaged leave 4.4 of nine policy questions “unsure”, against 1.0 for Advocates, and 27% say the word psychedelics brings nothing to mind. Only 28% feel negative. They are not against anything; they are not there.

19%

Research is the common ground.

Support for letting scientists study psychedelics runs from 50% to 88% across the five audiences, and opposition peaks at 19%, lower than the worst any other proposal faces. It has majority support in four of the five; among the Disengaged it reaches 50%, held back by 39% answering unsure rather than by opposition.

3×

After research, medicine is the opening with skeptics.

Opponents are more than three times as likely to back psychedelic therapy in a controlled medical setting (38%) as legal use for anyone over 21 (11%). The same gap runs through the middle: Curious Moderates give 68% to medical therapy and 44% to legal adult use.

3.4%

Party explains little.

Party identification explains just 3.4% of the differences in how much Americans support psychedelics policy, and age 6.1%. Republican identification runs only from 28% to 42% across the five audiences.

02 · Why segment

Why one number isn’t enough

Psychedelics have moved from the counterculture onto the policy agenda, and the public’s signals point in different directions at once.

In August 2024 the FDA declined to approve MDMA-assisted therapy for PTSD and asked for another late-stage trial. Three months later, Massachusetts voters rejected a measure to legalize natural psychedelics, 56% to 43%. In April 2025 New Mexico legalized psilocybin for medical use, the third state to allow legal access and the first to do it through its legislature rather than a ballot measure. A year later, the FDA put a psilocybin therapy for treatment-resistant depression on its fastest review track.

A single topline can’t explain that pattern. Ask Americans about psychedelics and the answer depends heavily on the question.

Support depends on what you ask

All US adults, percent on each of nine proposals

How much do you support or oppose the following? Five-point scale from strongly oppose to strongly support, with “unsure” as the middle option. Support and oppose combine the two points on each side. Weighted, n = 1,525.

Support runs from 69% for letting scientists study psychedelics down to 29% for letting drug companies patent their own versions, and on every proposal between about a fifth and about a third of Americans say they are unsure. Those numbers tell you where the public is. They don’t tell you who can still be moved, what would move them, or who they would listen to.

That is what a segmentation is for. Instead of one average American, it finds the handful of distinct audiences that actually make up the public, each with its own mix of beliefs, experience and trust. Strategy can then be built audience by audience, instead of for a person who doesn’t exist.

03 · Our approach

Nine questions, five audiences

We surveyed 1,527 US adults in November 2023 and weighted the results to represent the adult population. Nine of their answers were enough to sort them.

From answers to audiences. The nine questions produce seven signals. A statistical model finds the groups; policy and trust questions are held back to test them. Shares weighted, n = 1,525.

Compare, don’t just count.

Some people answer every question enthusiastically; others never go past “moderately.” So we compare each person’s answers with each other: how useful psychedelics are as therapy against how harmful they are to mental health, and so on. Rating both sides of a pair equally is real information. Rating everything high is not a pro-psychedelics view.

Let the data draw the lines.

A method called latent class analysis looks for groups of people whose answers follow the same pattern, without being told what the groups should be. We tested eleven candidate designs at three to six groups to choose the questions, then fitted the chosen design at two through seven to choose the number of audiences, taking the point that gives distinct, stable, useful groups from the fewest questions.

Test on questions it never saw.

No policy or trust question went into the model that forms the audiences, so how the audiences differ on them is evidence the groups are meaningful rather than arbitrary. The five audiences account for 39% of the differences in policy support, against 3.4% for party identification. (We did compare candidate designs on these measures when choosing one, so they are held out from the model, not from the choice.) Everyone who answered the nine questions is included.

04 · The audiences

Meet the Psychedelic Five

Two audiences have settled views at opposite ends. The three in between look alike in a poll average and want completely different things.

Where each audience sits

Each small dot is one respondent; each circle is an audience, sized by its share of US adults. Click a circle to jump to its profile.

Horizontal: benefits minus harms, averaged over three pairs rated by the same person (useful as a mental health therapy vs. harmful to mental health; good for society vs. causes violent crime; useful for treating addiction vs. addictive), each rated from “not at all” to “extremely.” Vertical: “How familiar are you with the topic of psychedelics?” Dots are nudged slightly so overlapping answers stay visible. Circles mark each audience’s weighted average; dashed lines mark the US average. n = 1,525.

The middle is three audiences, not one.

Conflicted Insiders, Curious Moderates and the Disengaged all sit near the national average on benefits versus harms, which is why a single number hides them. They have almost nothing else in common. Conflicted Insiders know the subject well (85% very or extremely familiar, 82% have tried one) and are unconvinced: only 22% see more benefit than harm. Curious Moderates know almost nothing (5% familiar) and already lean favourable: 54% see more benefit than harm. The Disengaged know almost nothing and have not formed a view at all: they leave 4.4 of nine policy questions unsure and 27% say the word brings nothing to mind. One needs reassurance, one needs a reason to commit, one needs to hear about the subject at all.

Convinced and experienced

Advocates

Advocates have made up their minds, and most say they have firsthand experience. Almost all of them (98%) see more benefit than harm in psychedelics, 93% feel positive about them, and 81% say they have tried one. They know the landscape: on average they have heard of 5 of the eight substances we asked about, more than any other audience.

98%see more benefit than harm

74%back legal, regulated adult use

40%are under 35 (US: 29%)

First word that comes to mindMushrooms 26% Tripping 15% LSD or acid 10%

Read the playbook →

Close to it, and torn

Conflicted Insiders

By their own account, Conflicted Insiders are the most familiar with psychedelics of any audience: 85% call themselves very or extremely familiar and 82% say they have tried one. They feel warmly toward psychedelics (78% positive), yet when asked about specific benefits and harms they call it roughly even: only 22% see more benefit than harm. They are the youngest audience and the only one more willing than Advocates to let drug companies patent psychedelics.

85%very or extremely familiar

22%see more benefit than harm, though 78% feel positive

56%back drug-company patents (US: 29%)

First word that comes to mindFun or good 14% LSD or acid 13% Mushrooms 11%

Read the playbook →

Warm, and new to it

Curious Moderates

Curious Moderates are the largest audience that has not settled. Only 5% call themselves very or extremely familiar and 79% say their feelings are neutral, yet asked about specific benefits and harms, 54% come down on the benefit side. They already back research (74%) and therapy in a medical setting (68%), and they leave 2.7 of nine policy questions unsure, against 4.4 for the Disengaged.

54%see more benefit than harm, on 5% familiarity

68%back therapy in a medical setting

79%call their feelings neutral

First word that comes to mindLSD or acid 18% Mushrooms 15% Nothing or not sure 12%

Read the playbook →

Not in the conversation

Disengaged

The Disengaged are not opposed so much as absent. 68% feel neutral, 66% say they are not at all or only a little familiar, and they pick the middle answer more than any other audience. They leave 4.4 of the nine policy questions unsure, more than four times as many as Advocates, and their most common first reaction to the word psychedelics is having none at all.

4.4 of 9policy questions left “unsure”

66%know little or nothing about the topic

27%first word is nothing, or not sure

First word that comes to mindNothing or not sure 27% Mushrooms 14% Drugs 8%

Read the playbook →

Settled against

Opponents

Opponents are the largest audience and the most certain in their beliefs. 86% feel negative about psychedelics, virtually all of them weigh the harms above the benefits, and they give the fewest middle answers of any audience. They skew older (36% are 65 or over, against 22% nationally) and somewhat more Republican, but they are not a partisan bloc. Their one clear opening is research.

86%feel negative about psychedelics

36%are 65 or older (US: 22%)

63%still back research

First word that comes to mindLSD or acid 20% Drugs 16% Danger or fear 16%

Read the playbook →

The five audiences side by side

Dashed lines mark all US adults. Hover any bar for the exact figure.

Feelings: “How positive or negative are your feelings towards psychedelics?” Familiarity: “How familiar are you with the topic of psychedelics?” Tried: reported ever trying any of ayahuasca, DMT, ibogaine, ketamine, LSD, mescaline, MDMA or psilocybin; self-reports in online surveys run high, so compare audiences rather than read these as national usage rates. More benefit than harm: the average of the three benefit-minus-harm pairs is above zero. Unsure: number of nine policy questions answered “unsure.” Policy items: percent who somewhat or strongly support, except where the label says unsure or oppose. Who they are: demographics, none of which was used to build the audiences. Weighted, n = 1,525.

“When you think of psychedelics, what is the first word that comes to mind?”

The five most common themes in each audience, percent of the audience

Open-ended answers grouped by keyword (for example, “shrooms” and “magic mushrooms” both count as Mushrooms; “dangerous,” “scary” and “illegal” as Danger or fear). An answer can match more than one theme; 25% of answers match none. Weighted, n = 1,525.

05 · Who they are

Age separates them. Party barely does.

No demographic was used to build the audiences, so everything on this page is something the audiences turned out to have rather than something they were built from.

Every audience against the national figure, group by group

Percent of each audience; dashed lines mark all US adults. Hover any bar for the exact figure.

Percent of each audience falling in each group. Columns within a block sum to 100 for each audience. Party counts leaners with the party they lean toward. Weighted, n = 1,525.

Age does real work.

Advocates and Conflicted Insiders are the youngest audiences, with 40% and 42% under 35 against 29% nationally. Opponents are the oldest: 36% are 65 or over against 22%. Age explains 6.1% of the differences in policy support.

Party does very little.

Republican identification runs only from 28% to 42% across the five audiences, and every audience contains large numbers of both parties. Party explains 3.4% of the differences in policy support, against 39% for the audiences themselves.

Education and race barely separate them.

Education explains 0.3% of the differences in policy support and race 0.3%, both close to nothing. The one visible tilt is that Conflicted Insiders are more likely to be Black than average, and that both Opponents and the Disengaged are more likely than average to be women (57% and 56% against 50%), with the Disengaged also the most likely to have a high school education or less.

Which is why you cannot buy these audiences on demographics.

No demographic filter isolates any of the five. The practical implication is that these audiences are reached by message and messenger rather than by targeting, except for the two age-tilted ends.

Everything on one scale

How much of the variation in policy support each way of grouping people accounts for

Weighted eta-squared on average support across seven policy propositions, the same measure for every row: the share of the differences between people that sits between groups rather than within them. The audiences were built without any of these questions. Weighted, n = 1,525.

06 · Try it yourself

Which one are you?

The whole segmentation runs on nine questions. Answer them exactly as our respondents did, and the same model that sorted 1,525 Americans will place you.

01In general, how familiar are you with the topic of psychedelics (e.g., LSD, psilocybin mushrooms)?

02In general, how positive or negative are your feelings towards psychedelics?

03–08To the best of your knowledge, to what extent are psychedelics…

	Not at all	Only a little	Moderately	Very	Extremely
Addictive
Harmful to people’s mental health
Useful as a mental health therapy
Useful for treating addiction
Good for society
Responsible for causing violent crime

Your answers stay on this page. Nothing is sent, stored or tracked.

You are most like

See how to reach this audience

Take the quiz in the interactive version of this report.

07 · Policy

Research unites. Legalization divides.

None of these questions was used to build the audiences, yet the audiences differ sharply on almost all of them.

“How much do you support or oppose the following policies?” Percent who somewhat or strongly support, answer “unsure,” or somewhat or strongly oppose; the three views are switched with the buttons above, and all three are printed in the PDF. Therapy items for PTSD and depression were asked separately, after a sentence noting that some research suggests each could be effective. Weighted, n = 1,525.

Research is the opening line.

Nobody organises against it: opposition to letting scientists study psychedelics runs from 5% to 19% across the five audiences, a narrower band and a lower ceiling than any other proposal faces. It draws majority support in four of the five and 50% among the Disengaged, where the shortfall is 39% unsure rather than opposition.

Medicine outpolls legalization.

Nationally, 58% back therapy in a controlled medical setting and 37% back legal use for anyone over 21. The gap runs through every audience and is widest among Opponents: 38% against 11%.

“Unsure” is the story in the middle.

Switch to “Unsure” above: the Disengaged column runs ahead of every other on every item. On MDMA therapy for PTSD, 55% of the Disengaged have not decided, against 37% of Curious Moderates and 32% of all adults. Curious Moderates are unsure more often than the country too, but on a different scale. Splitting the old middle in two separates the people who have a view they have not voiced from the people who have no view yet.

Commercialization is the hardest sell.

Only 29% of Americans back letting drug companies patent and sell psychedelics. Conflicted Insiders are the one audience that backs it more than Advocates do (56% against 51%), and the Disengaged are the least keen of the three middles.

08 · Messengers

Clinical voices travel. Lived experience divides.

Who each audience trusts as a source of information about psychedelics: how many trust each source outright, and which sources each audience ranks above its own average, relative to their average source.

AdvocatesConflicted InsidersCurious ModeratesDisengagedOpponents

“How much do you trust or distrust the following as a source of information about psychedelics?” Relative view: trust in each source minus that person’s average trust across all thirteen sources, averaged within each audience, in points on the five-point trust scale; right of zero means trusted more than that audience’s average source. This removes the fact that some audiences trust everything more than others do. Percent view: the share of each audience that somewhat or strongly trusts the source, with the dashed mark showing all US adults; use this to compare reach, and the relative view to compare priority. The grey bar spans the five audiences. Weighted, n = 1,525.

Doctors, scientists and psychiatrists are trusted everywhere.

All five audiences trust them above their average source. People who have had psychedelic therapy are the only other source to clear that bar in every audience, and only barely with Opponents. Relative to their other sources, Opponents lean on doctors and scientists more than any other audience does.

Lived-experience voices split the audiences.

People who have used psychedelics, experienced guides and Indigenous practitioners sit near the top for Advocates and below average for Opponents. In the middle they separate the two unfamiliar audiences: Curious Moderates give people who have had psychedelic therapy +0.27 and Indigenous practitioners +0.27, while for the Disengaged every one of them sits within a sixth of a point of their own average. First-person accounts have somewhere to land in one of those audiences and nowhere to land in the other.

Politicians rank last with everyone.

Every audience trusts politicians and elected officials less than any other source on the list.

Institutions split the base from the opposition.

The FDA and law enforcement rank clearly above average only with Opponents, and well below it with Advocates. Pharmaceutical companies are least unwelcome among Conflicted Insiders.

09 · Strategy

How to reach each audience

Our recommendations, drawn from what each audience believes, supports and trusts. The numbers come from the survey; the strategy is XandY’s interpretation of them.

Mobilize

Advocates

Turn conviction into action.

17% of adults

Why: Advocates are already persuaded. Persuasion dollars spent here buy little; what they can offer is energy, stories and credibility with each other.
Lead with: Give them something to do: share a story, show up for a hearing, recruit a friend. Frame the cause around healing and access, which they already believe in.
Messengers: People with lived experience, Indigenous practitioners and scientists. Four experiential voices sit level at the top: Indigenous practitioners (+0.64), people who have had psychedelic therapy (+0.62), experienced guides (+0.62) and people who have used psychedelics (+0.61), all well above their own average source.
Avoid: Industry and government voices. Advocates rank politicians (-1.45), law enforcement (-1.19) and pharmaceutical companies (-0.91) far below their average source. And keep their favorite messengers inside Advocate channels: experienced guides and Indigenous practitioners rank below average with Opponents.
Where to find them: Younger and more male than average: 40% are under 35. Existing networks, communities and organizations they already belong to.

Reassure

Conflicted Insiders

Resolve the conflict with safety, not enthusiasm.

12% of adults

Why: Conflicted Insiders have seen both sides up close. Telling them psychedelics are wonderful restates half of what they already believe and ignores the other half.
Lead with: Take the risks seriously in the first sentence, then show the guardrails: screening, dosing, trained supervision, tested products, licensed settings. “Done right” is their frame. They are the audience for regulated-market models, and the only one where drug-company patents draw more support (56%) than among Advocates.
Messengers: Doctors (+0.28), experienced guides (+0.26) and scientists (+0.25) are level at the top of their list: people who know what doing it well looks like. They are also the audience least put off by pharmaceutical companies (-0.16, against -0.44 nationally).
Avoid: Pure enthusiasm, and politicians, who rank lowest with them as with everyone.
Where to find them: The youngest audience: 42% are under 35 and only 3% are 65 or older. More likely than average to be Black (20% against 10%). Digital channels and peer settings over broadcast.

Convert

Curious Moderates

Turn a warm lean into stated support.

22% of adults

Why: This is the biggest movable audience in the country and the shortest distance to travel. 54% already judge the benefits to outweigh the harms on the specifics, and 68% back therapy in a medical setting, but only 11% describe their overall feelings as positive. The gap between what they think item by item and what they feel in general is the opening.
Lead with: Name what they already half believe and give it a label. Concrete clinical cases, real patients, trial results, and what supervised therapy actually looks like. They do not need to be argued out of an objection; they need permission to say out loud what the specifics already tell them.
Messengers: Scientists (+0.54), doctors (+0.49) and psychiatrists (+0.48). They also give people who have had psychedelic therapy (+0.27) and Indigenous practitioners (+0.27) a clearer positive than either other middle audience does, so first-person accounts have more room to land here than anywhere outside the Advocates.
Avoid: Politicians (-1.12) and pharmaceutical companies (-0.42), and any framing that makes support a partisan position. Legalization language costs more than it gains: support falls from 68% for medical therapy to 44% for legal adult use.
Where to find them: Close to the national profile and hard to buy on demographics alone: 33% are under 35 against 29% nationally, and 28% lean Republican against 35%. Health and science media, documentary and long-form, mainstream news.

Reach

Disengaged

Get on the radar at all.

12% of adults

Why: The Disengaged are not arguing with anyone. They leave 4.4 of nine policy questions unsure and 27% say the word psychedelics brings nothing to mind. The barrier is attention, not opposition, and it is the most expensive barrier to clear.
Lead with: One idea, repeated, in plain words: doctors are testing psychedelics as a treatment for depression and PTSD, in hospitals, under supervision. Nothing about policy, nothing about legalization, no numbers. If they hear one sentence about this topic in a year, make it that one.
Messengers: Doctors (+0.34), scientists (+0.32) and psychiatrists (+0.21), and nobody else. Every experiential voice sits within a sixth of a point of their own average: people who have had psychedelic therapy +0.14, Indigenous practitioners +0.01, experienced guides -0.03. Testimonial has no purchase here.
Avoid: Anything that assumes prior knowledge, and politicians (-0.64). Also avoid reading their silence as agreement: on the specifics only 12% come down on the benefit side, well below Curious Moderates.
Where to find them: More female (56% against 50% nationally), less educated (48% have a high school education or less against 40%) and less partisan. Broad-reach television, radio and social feeds rather than anything topic-specific, because they are not seeking this subject out.

Bridge

Opponents

Keep opposition from hardening against research and medicine.

36% of adults

Why: Converting Opponents is not a realistic goal: 86% feel negative and they give the fewest middle answers of any audience. But 63% back research, and they are more than three times as likely to support therapy in a controlled medical setting (38%) as legal use for anyone over 21 (11%). Medical uses are where their opposition is softest after research.
Lead with: Stay on rigor: clinical trials, medical supervision, safety data, FDA oversight. Talk about studying psychedelics, not using them.
Messengers: Doctors (+0.76) and scientists (+0.69) above all. They are also the only audience that ranks the FDA (+0.29) and law enforcement (+0.25) clearly above its average source.
Avoid: Legalization and recreational framing, and experiential messengers: experienced guides (-0.34) and Indigenous practitioners (-0.22) rank below their average source.
Where to find them: Older: 36% are 65 or over, against 22% nationally. They lean Republican more than average (42% against 35%), but most are not Republicans, so partisan targeting alone will miss them.

RULE 01

Lead with research, then medicine.

No proposal is less objectionable at its worst: opposition to research peaks at 19%, among Opponents, and every other proposal peaks higher. It holds majority support in four of the five audiences. After research, medical uses draw the least opposition from Opponents: about a third oppose therapy for depression or PTSD, against 69% for legalization.

RULE 02

Put clinicians out front.

Doctors, scientists and psychiatrists are trusted above average by all five audiences. Only one other source manages that, people who have had psychedelic therapy, and only barely with Opponents. For the Disengaged the three clinical voices are the only sources that stand out at all.

RULE 03

Match the messenger to the audience.

Several of Advocates’ most trusted voices, including guides and Indigenous practitioners, rank below average with Opponents and barely register with the Disengaged. Aim them at the base, at Curious Moderates and at Conflicted Insiders rather than broadcasting them.

10 · About this research

How we did it

1,527US adults surveyed online, November 17 to 19, 2023

±3.1point margin of error for all adults, at 95% confidence

0.82classification certainty (entropy): how cleanly people sort into one audience

0.88average probability that a respondent belongs to their assigned audience

The survey

The XandY Omnibus surveyed 1,527 US adults online from November 17 to 19, 2023. Results are weighted to represent the US adult population; weighting widens the margin of error to ±3.1 points for all adults (design effect 1.49). Estimates within a single audience have wider margins, because each audience is a fraction of the sample; the ± figures on the audience profiles are margins on the size of each audience. 1,525 respondents answered all nine segmentation questions and are described here; nobody was excluded for any other reason. The survey was fielded before the policy events described in section 02, so these audiences describe the public as that period began.

The model

A five-class latent class model estimated in R (poLCA) from seven indicators built from nine questions: feelings toward psychedelics, familiarity, three benefit-minus-harm comparisons, the number of eight substances tried, and the number of middle answers on the six belief items. Policy, trust and demographic questions were never used to build the audiences. The number of audiences was chosen by fitting two through seven and taking the point where the fit statistics stop improving materially and the smallest audience is still large enough to brief: adding a sixth buys almost nothing and produces a group of five percent. The solution reproduces its own groupings at an adjusted Rand index of 0.80 across forty split-half refits. A companion technical report sets out every robustness check.

How sure are we

Every respondent is kept. Nobody is excluded for straightlining, for inconsistent answers or for any other response pattern: a segmentation that drops awkward respondents becomes a segmentation of cooperative survey-takers. 14% of adults gave at least one answer that does not add up, such as reporting having tried a substance they said they had never heard of, or giving the identical answer to all nine belief items.

Those answers are not spread evenly, and two audiences carry most of them: 44% of Conflicted Insiders and 43% of the Disengaged, against 9% of Advocates, 6% of Curious Moderates and 1% of Opponents. Weighted, 72% of the 97 respondents who gave one answer to all nine belief items sit in the Disengaged. Refitting the model without them recovers a recognisably similar grouping (adjusted Rand index 0.69) but not an identical one.

What follows from that: the Disengaged is a real audience, and its absolute figures are the least reliable in this report. Among its members who answered consistently, support for research is 69% rather than 50% and the number of policy questions left unsure is 3.1 rather than 4.4. Read that audience as low-information and low-engagement, which both versions agree on, rather than as the precise percentages. The same caution applies to the line between Curious Moderates and the Disengaged: a refit without the flat answers keeps 38% of Curious Moderates in place and draws that boundary in a different spot, so anyone planning against those two should be ready to treat them as one. Conflicted Insiders carry a similar caveat: among those who answered consistently, 85% are very or extremely familiar, 74% have tried one, and their benefits-minus-harms average is -0.38. Self-reported use in online surveys runs higher than in federal surveys; we use it to compare audiences, not to estimate national usage.

Sources for section 02

FDA complete response letter to Lykos Therapeutics, August 9, 2024: AJMC
Massachusetts Question 4, November 2024, failed 56% to 43%: Axios Boston
New Mexico Medical Psilocybin Act, signed April 7, 2025: Foley & Lardner
FDA rolling review and national priority voucher for COMP360 psilocybin, April 24, 2026: Compass Pathways

Reproducibility

Every number in this report is produced by R scripts that run from the raw survey file: PSY5F_pipeline.R estimates and freezes the model, and PSY5F_public_numbers.R computes the figures shown here. The quiz uses the frozen model and reproduces its assignment for all 1,525 respondents.