Category: Science Explained
The placebo effect is not people imagining things. It is measurable improvement produced by expectation, and it is largest for exactly the outcomes supplements are sold on: sleep, mood and pain. Here is why that makes proper blinding the difference between a real result and a flattering one.
Category: Science Decoded | Reading time: ~9 min | Level: Intermediate
A friend starts a new sleep supplement and, three nights later, tells you it changed everything. They mean it. They really are sleeping better, and their eyes are clearer, and they are grateful. And none of that, on its own, tells you whether the supplement did anything at all. That is the uncomfortable, fascinating heart of the placebo effect: people can genuinely improve on a treatment that contains nothing active, and they are not fooling themselves in any simple sense.
This is not a trick or a footnote. It is one of the most reliable findings in all of medicine, and it lands hardest on exactly the products this industry sells: things for sleep, calm, mood and pain. Understanding the placebo effect is the single most useful skill for reading supplement claims honestly, because it explains why feeling better after taking something is such weak evidence, and why the whole apparatus of blinded trials exists to get past it.
The placebo effect is a real, measurable improvement in symptoms that follows an inactive treatment, such as a sugar pill or a saline injection, driven not by the treatment's contents but by everything around it [1]. Expectation is the engine: believing you are being helped, the ritual of taking something, the attention of a clinician and the natural tendency of many symptoms to ease over time all feed the response [1][3].
The mechanism worth remembering is that this is not imagination in the dismissive sense. Studies show placebo responses can involve genuine changes in brain activity and physiology, and the improvement people report is real to them [1]. What is not established, in a placebo response, is that the treatment itself caused it. That distinction, real improvement, but not from the compound, is the whole reason the effect matters.
The placebo effect has a shadow. When people expect a treatment to harm them, they can develop real, unpleasant symptoms from something inert, and this is called the nocebo effect [3]. In trials it shows up constantly: participants given a dummy pill frequently report headaches, nausea or fatigue, often the very side effects they were warned about during consent [3].
This matters for two reasons. It shows that expectation cuts both ways and is powerful enough to manufacture symptoms, not just relief. And it is a caution against assuming that a side effect after taking a supplement proves the supplement caused it, just as a benefit does not prove efficacy. The framing of information, the warnings, the anecdotes, the marketing, shapes how people actually feel.
Here is the part that bears directly on supplements. The placebo effect is not equally large everywhere. For objective, hard-to-influence measures, blood cholesterol, tumour size, a fracture healing, placebo effects tend to be small [2]. For subjective, self-reported experiences, pain, sleep quality, anxiety, mood, fatigue, they can be large, sometimes accounting for a substantial share of the improvement a treatment group shows [1][2].
Now look at what supplements are sold on. Better sleep. Calmer mood. Less anxiety. More energy. A clearer head. These are almost entirely subjective outcomes, the exact category where the placebo response is strongest [2]. Add the ideal conditions for a placebo, a person who chose to buy the product, expects it to work, and takes it as a comforting nightly ritual, and you have a powerful improvement engine running before any active compound does a thing.
This does not mean supplements are useless. It means that in an uncontrolled setting, a herbal remedy for sleep is competing against, and easily confused with, a placebo response that is naturally big for sleep. Feeling better is not the evidence people think it is.
Two further forces stack on top of the placebo response and are easily mistaken for it. The first is regression to the mean: people tend to start a remedy when their symptoms are at their worst, and symptoms that are unusually bad often improve on their own simply because they were at a peak. The second is the natural waxing and waning of most complaints, sleep and mood among them, which drift up and down regardless of treatment. Neither is the placebo effect strictly, but both push an uncontrolled before-and-after towards looking like a benefit, and only a comparison group experiencing the same drift can tell them apart from a real effect.
If the placebo effect can mimic efficacy, how does anyone ever know a treatment truly works? The answer is the placebo-controlled, double-blind trial, and it is one of the most important inventions in medicine [4].
A placebo-controlled trial gives one randomly assigned group the active treatment and another an identical-looking dummy, then compares them [4]. Both groups experience the placebo response, so whatever benefit the active group shows over and above the placebo group is the part attributable to the treatment itself. If the active treatment does not clearly beat the placebo, the honest reading is that its apparent benefit was mostly placebo.
Blinding makes this work. In a double-blind trial, neither the participants nor the researchers interacting with them know who got what until the end [4]. This blocks two distortions at once: participants who know they got the real thing inflate their reported improvement, and researchers who know can unconsciously nudge results and interpretation. A trial that is not blinded, an open-label study where everyone knows who took the active supplement, is wide open to exactly the inflation the placebo effect produces, and its glowing results should be read with real caution [2][4].
The practical skill is to downgrade certain kinds of evidence in your own head. A friend's testimonial, a five-star review, your own rapid improvement on a calming supplement, these are all real experiences and all weak evidence of efficacy, because the placebo effect is large for precisely these outcomes and fast to appear.
When you assess a remedy, ask a sharper question than did people feel better. Ask whether it beat a placebo, in people who did not know which they were taking, over a sensible period. If the only support is uncontrolled testimonials or an open-label study, treat the claim as unproven, however sincere the enthusiasm behind it. And notice that this cuts both ways for side effects too: the nocebo effect means anxiety about a product can generate real discomfort, so alarming anecdotes deserve the same scepticism as glowing ones. Our piece on how we grade evidence puts this into a full framework, and our honest look at whether valerian root actually works shows the placebo problem in one worked example.
The placebo effect raises an ethical line as much as a scientific one. It is legitimate to acknowledge that a calming routine, positive expectation and the context of care genuinely help people feel better. It is not legitimate to claim a specific ingredient is responsible for benefits that are really the placebo response, or to manufacture expectation in order to sell an inert compound at a premium.
The honest position is to credit a compound only for what survives a proper placebo comparison, and to be candid that ritual and expectation are doing real work alongside, or sometimes instead of, the active ingredient. That candour is not a weakness in a remedy's story. It is what makes the story trustworthy.
Pregnant, breastfeeding, or on medication? Check with a healthcare professional first.
We treat the placebo effect as central to honest grading, not as an inconvenience. When we describe a remedy, the quality of its placebo control shapes the grade we give it, which is why a well-blinded null result can weigh more with us than an enthusiastic uncontrolled study. Our aim is to credit ingredients for real effects and to be plain when the apparent benefit is mostly expectation.
If you want the full framework, our explainer on how we grade evidence lays out the tiers, and our piece on whether valerian root actually works applies them to a popular sleep herb where the placebo problem is front and centre. You can also browse the Remedy Library to see how we describe what an ingredient does, and what it does not, once expectation is accounted for.
1. Finniss DG, Kaptchuk TJ, Miller F, Benedetti F (2010). Biological, clinical, and ethical advances of placebo effects. The Lancet, 375(9715), 686 to 695. Review of placebo mechanisms including expectation and physiological change. 2. Hrobjartsson A, Gotzsche PC (2010). Placebo interventions for all clinical conditions. Cochrane Database of Systematic Reviews, (1), CD003974. Systematic review finding larger placebo effects for subjective, patient-reported outcomes such as pain and nausea. 3. Colloca L, Barsky AJ (2020). Placebo and nocebo effects. New England Journal of Medicine, 382(6), 554 to 561. Review of placebo and nocebo mechanisms and clinical relevance. 4. US Food and Drug Administration (2023). Randomised, placebo-controlled and blinded trials. Regulatory guidance on the role of placebo control and blinding in demonstrating effectiveness.
It is a genuine improvement in how someone feels or functions after receiving a treatment that has no active ingredient, such as a sugar pill. The improvement is driven by expectation, the ritual of being treated, the attention of a clinician and the natural tendency of many symptoms to ease over time. Crucially it is not people lying or imagining things. Brain-imaging and physiological studies show real changes can accompany a placebo response. It is one of the most reliable findings in medicine, which is exactly why trials have to control for it.
The nocebo effect is the placebo effect's opposite. When someone expects a treatment to cause harm or side effects, they can experience real, unpleasant symptoms even from an inactive treatment. In trials, people given a dummy pill often report side effects such as headache or nausea, simply because they were warned those might occur. The nocebo effect explains why anxiety about a supplement or medication can produce genuine discomfort, and why the way information is framed influences how people feel.
Because most supplements are sold on subjective outcomes, better sleep, calmer mood, less anxiety, more energy, and these are precisely the outcomes where the placebo effect is strongest. A person who buys a calming supplement, expects to relax and takes it as a nightly ritual has several powerful placebo ingredients working before any active compound does anything. That does not prove supplements do nothing, but it does mean that feeling better after taking one is weak evidence on its own, and that only properly controlled trials can separate the compound from the expectation.
It is a study that compares an active treatment against an inactive dummy, the placebo, that looks, tastes and feels identical. Participants are randomly assigned to one or the other. By comparing the two groups, researchers can see how much benefit is due to the treatment itself over and above the placebo response that both groups experience. If the active treatment does not clearly beat the placebo, the honest conclusion is that its apparent benefit was mostly placebo, not the compound.
Blinding means keeping people unaware of who received the real treatment and who received the placebo. In a single-blind trial the participants do not know; in a double-blind trial neither the participants nor the researchers interacting with them know until the study ends. Blinding matters because knowing you got the active treatment inflates your reported improvement, and researchers who know can unconsciously influence results. Double-blinding is the standard defence against both, and a trial that is not blinded should be read with real caution.
No. It means you cannot tell whether a remedy works from the fact that people feel better after taking it, because some of that improvement would have happened with a dummy pill. Plenty of remedies have real effects that survive placebo-controlled testing, and plenty do not. The placebo effect is not an argument against herbs, it is an argument for judging them by controlled trials rather than by testimonials, which is the whole basis of honest evidence grading.
It varies a lot depending on the condition and the outcome. For objective, hard-to-fake measures such as blood cholesterol or tumour size, placebo effects tend to be small. For subjective, self-reported experiences such as pain, sleep quality, mood and fatigue, they can be substantial, sometimes accounting for a large share of the improvement seen in a treatment group. That gap is why the same supplement can look impressive in an uncontrolled setting and unremarkable once tested against a proper placebo.
It is real that you felt calmer, but on its own it is weak proof the compound caused it. Calm and sleep are subjective outcomes where the placebo effect is large, and expectation, routine and the simple act of doing something about a problem all contribute. This is not a reason to dismiss your experience, it is a reason to trust controlled trials over personal before-and-after for deciding whether an ingredient genuinely works. The honest question is not did I feel better, it is did this beat a dummy pill in people who did not know which they took.
Because blinding is what separates a real effect from a flattering one, especially for the sleep, mood and stress claims supplements lean on. An unblinded or uncontrolled study can report a big improvement that is mostly placebo, and treating that as efficacy would mislead readers. When we grade a remedy, the quality of its placebo control is part of the grade, which is why a well-blinded null result can count for more than an enthusiastic uncontrolled one.
To a degree, yes. The context of care, a calming routine and positive expectation genuinely help people feel better, and there is nothing dishonest about acknowledging that. The line is crossed when a product claims its specific ingredient is responsible for benefits that are really the placebo response, or when marketing manufactures expectation to sell an inert compound. Being honest means crediting the compound only for what survives a proper placebo comparison, and being clear that ritual and expectation are doing real work too.