The label on an experimental cigarette can become part of the experiment. Give two groups the same placebo and tell only one that it contains THC. If they subsequently describe the experience differently, something besides the preparation has changed. Now repeat the comparison with active THC. These four possibilities make it possible to investigate a question that ordinary placebo-controlled trials usually leave in the background. What does somebody bring to the experience simply by expecting a drug?
Four groups, two questions
In a pilot study led by Jane Metrik, twenty cannabis smokers took part in what researchers call a balanced placebo design. Some people received the preparation the experimenter said they would receive. Others heard that the preparation contained THC when they received placebo, or heard that it contained no THC when it did. The design crossed the information with the substance so that each could vary independently.
Compare two groups given placebo and any consistent difference between them cannot result from a difference in administered THC. Compare groups given identical instructions but different preparations and the question changes. It now concerns the effect of the substance when the initial message is held constant. A further possibility is an interaction: perhaps the instruction matters more when bodily sensations leave room for interpretation than when the drug’s effects are easy to recognize.
Why ask about the cigarette itself?
In the small pilot, the instructions influenced how people rated features including potency, smell, taste and satisfaction. This is striking because an experimenter’s words cannot alter the molecules already present in a cigarette. They can influence what a person attends to, how they interpret a familiar sensation or which point they choose on a rating scale. The pilot gives reasons to investigate those possibilities. With twenty participants it gives little basis for predicting how large any one effect will be in a wider population.
An instruction is also an intervention in its own right. If somebody expects impairment, they might take a task more slowly or check an answer twice. Someone expecting relaxation could pay special attention to changes in tension. A study of expectations about impairment examined how beliefs about a drug’s effects related to risk taking, an outcome that depends on choices as well as on an ability to act. Describing either response as an illusion would miss the behavior that actually occurred. The more interesting question is which observable outcomes changed along with the report.
Feeling intoxicated and stopping a response
A larger experiment involving 136 regular users used the same four-way design. Both receiving THC and expecting THC increased subjective intoxication ratings independently. On tasks requiring participants to interrupt or inhibit a response, actual THC impaired performance regardless of the instruction. The two outcomes told different stories in the same participants. A reported feeling could shift with the information supplied, while aspects of response control remained sensitive to the drug.
People can also revise a belief during a session. Someone told placebo may begin to suspect otherwise. A review of blinding in cannabinoid trials explains why recognizable drug effects complicate interpretation. Simply removing participants who voice doubts about the message introduces a selection problem: sensitivity to the drug or familiarity with its effects could be among the reasons those people doubted it.
What the volunteer actually hears
A broader systematic review of context and subjective cannabis effects places these laboratory instructions among other features of the setting. A sentence delivered by an experimenter is a particular kind of expectation. A person arriving with a long history of cannabis use already has beliefs about the drug, and the presence of laboratory staff may give the words unusual authority.
That leaves the volunteer who expected THC and received placebo with a perfectly coherent experience to report. A familiar flavor may seem stronger; a quiet moment may acquire significance. The decisive detail is which question the researcher asks afterward. A rating of potency, an account of feeling intoxicated and a failed attempt to stop a response need not move together. Asking all three gives expectation a measurable place in the story without assigning it every effect in the room.