No. NLP as a whole is not scientifically proven: the research base is small, methodologically weak, and the reviews that exist have mostly found insufficient evidence for its core claims. Some individual components overlap with techniques that are well evidenced elsewhere — exposure, cognitive reframing, goal specification, rapport-building — but the overlap is not the same as NLP itself having been validated.
That is the honest answer, and it is worth giving before the defence. Neuro-Linguistic Programming is a model of communication and change built in the 1970s by modelling what effective therapists did; it was never built as a scientific theory, and it has not been tested like one.
| Question | Short answer |
|---|---|
| Is NLP scientifically proven? | No. The evidence base is thin and largely negative or inconclusive |
| Has it been tested? | Yes, mainly in the 1980s, plus a few later reviews |
| What failed testing? | Eye accessing cues as lie detection; matching representational systems for rapport |
| What is better supported? | Elements shared with other methods: exposure, reframing, specific goals |
| Why is it controversial? | Overclaiming, no regulation, weak research, commercial incentives |
| Does that make it useless? | No — but “unproven” is the accurate word, not “proven” |
| Is it a therapy? | No. It complements professional care and does not replace it |
What does the research on NLP actually say?
The reviews are consistent: not enough good studies, and what exists rarely supports the strong claims. Tomasz Witkowski’s 2010 analysis in the Polish Psychological Bulletin examined the published NLP literature and concluded that the substantial majority of the studies he could assess did not support NLP’s hypotheses. Two years later, a systematic review by Sturt and colleagues in the British Journal of General Practice looked specifically at NLP’s effect on health outcomes and found the available evidence insufficient to draw conclusions either way.
“Insufficient evidence” is not the same finding as “proven ineffective”, and both get misquoted by the side they suit. What the reviews describe is a field that has largely not been studied to the standard required — small samples, missing control groups, outcomes measured by the people delivering the intervention, and very little replication.
| Claim | Status of the evidence |
|---|---|
| Eye movements reveal whether someone is lying or recalling | Tested and not supported. Wiseman and colleagues’ 2012 experiments in PLOS ONE found no reliable link |
| Matching someone’s “preferred representational system” builds rapport | Tested repeatedly in the 1980s; largely unsupported |
| The fast phobia process reduces a specific phobic response | Weak, small-scale evidence; the underlying mechanism resembles well-evidenced exposure work |
| Reframing changes the meaning attached to an event | Not proven as an NLP technique, but consistent with well-evidenced cognitive approaches |
| Specific, sensory-defined goals outperform vague intentions | Supported — though by goal-setting research, not by NLP research |
| NLP treats trauma, depression or anxiety disorders | Not supported. This is a clinical claim NLP is not entitled to make |
The pattern in that table is the honest summary of the field: the modest, mechanical claims hold up best, and the dramatic ones — the ones that sell seminars — hold up worst. Eye accessing cues are the clearest example: still taught in some courses as a lie-detection tool, still unsupported by the research.
Why is NLP controversial?
Because the gap between what NLP is marketed as and what it can demonstrate is unusually wide, and because nothing in the field forces that gap to close. Four things drive the controversy, and all four are legitimate criticisms rather than misunderstandings.
- Overclaiming. “Cure a phobia in ten minutes” and “reprogram your brain” are sales lines that no evidence supports. They also make it harder to discuss the parts that do work.
- Scientific-sounding language without the science. “Neuro” and “programming” imply a neurological model that NLP does not have. Critics call this borrowed authority, and they have a point.
- No regulation and no gatekeeping. Anyone can run a training, issue a certificate and invent an association. Quality varies enormously between two people holding the same-looking document.
- Commercial incentives against testing. Rigorous research is expensive and risks producing an unwelcome answer. A field funded by seminar income has little structural reason to commission it.
There is a fifth, quieter reason: the founders’ own style. NLP grew from workshops and demonstrations rather than journals, and demonstration is persuasive in a way that data is not. A room watching one dramatic change in ten minutes will believe more than a study can justify — which is precisely why the field needs the studies it does not have.
Does unproven mean it does not work?
No, but it does mean any claim about effectiveness should be modest and testable in your own case. Absence of good evidence is a statement about the research, not a verdict on every individual outcome — and it is also not permission to assume the outcomes are real.
Several ordinary explanations account for a good deal of what people experience in NLP work, and none of them require the model to be true: focused attention on a problem you have avoided, a clear and specific goal, an hour of undivided attention from someone skilled at asking questions, expectation, and gradual exposure to something previously avoided. Those factors improve outcomes across most talking interventions. If NLP has a defensible claim, it is that it packages some of them efficiently — not that it has discovered a separate mechanism.
This matters practically when you decide what to spend money on. A training that says “these are structured tools for communication and self-management, and here is what they cannot do” is describing reality. One that promises to rewire your neurology is selling something the evidence does not back. The same test applies to a practitioner: ask what they think NLP cannot do, and listen to whether the answer is specific. If you want the underlying model without the marketing, the plain explanation of what NLP is is a better starting point than any seminar page.
How should you evaluate NLP claims yourself?
Treat every claim as a prediction and check whether it can fail. Claims that cannot fail — “it works if you believe in it” — are the ones to discard first, whoever is making them.
- Ask what would count as it not working. A trainer who cannot answer has not thought about it, or does not want you to.
- Separate the technique from the mechanism. A process can be useful while its stated explanation is wrong. Both statements can be true at once.
- Check the source of a study. Research published by the organisation selling the training is a starting point, not evidence.
- Watch for medicalised language. “Treat”, “cure”, “heal” and “diagnose” are clinical words. NLP is not a clinical discipline.
- Test on something measurable. A specific behaviour, in a specific situation, with an outcome you would notice if it failed.
I teach this material and I still think the honest position is the uncomfortable one: the model is unproven, some of the early claims were plainly wrong, and the techniques are still worth learning as structured communication tools. What I can defend from the training room is narrow and observable — people become better at noticing what happens in a conversation before it goes wrong. That is not a scientific claim, and I would not present it as one. Trainings that promise more than that, including any of mine, deserve the same scrutiny as the research. — Carolin Mallmann, Licensed Trainer of The Society of NLP, Lisbon
If you want to judge the work rather than the argument about it, seeing techniques run live is more informative than any reading list — which is why the Practitioner training we run in Lisbon is built around supervised practice in small groups rather than lecture time. Come sceptical; it is the more useful state to learn in.
This article is informational and does not constitute medical or psychological advice. NLP work complements professional care and does not replace it. If you are dealing with a diagnosed condition such as an anxiety disorder, depression or trauma, speak with a licensed clinician.
Frequently asked questions
Is NLP considered a pseudoscience?
By much of academic psychology, yes — chiefly because it uses neurological language for claims it has not demonstrated, and because several of its early testable claims failed when tested. Practitioners often counter that NLP is a practical model rather than a scientific theory. Both positions can be held honestly; what cannot be defended is presenting NLP as scientifically established.
Are there any studies that support NLP?
Some small studies report positive results, particularly on anxiety symptoms and phobic responses, but they tend to have small samples, no adequate control condition, or authors with a commercial interest in the outcome. Systematic reviews weighing this literature as a whole have concluded that the evidence is insufficient rather than convincing.
Why do people say NLP works if it is not proven?
Because individual experience and scientific evidence answer different questions. A person can genuinely improve after a session through attention, expectation, a clearer goal or exposure to something previously avoided. Evidence is about whether the method outperforms those factors across many people — which is what has not been shown.
Is NLP recognised by the NHS or health authorities?
No. NLP is not a recommended treatment in mainstream clinical guidelines, and it is not a regulated healthcare profession in the UK or the EU. Practitioners may work on non-clinical patterns such as confidence, communication or specific fears, but they must not present that work as treatment for a diagnosed condition.
Do NLP trainers acknowledge the criticism?
The good ones do. A trainer who can name which claims failed testing, and which parts of the toolkit are shared with better-evidenced methods, is describing the field accurately. A trainer who dismisses all criticism as closed-mindedness is a reason to look elsewhere, regardless of their certification.
Should I avoid NLP because it is unproven?
Not necessarily — but calibrate what you expect and what you pay. Treated as structured communication and self-management tools with an unproven theoretical wrapper, NLP can be worth learning. Treated as a substitute for therapy, or bought on the promise of rapid transformation, it is a poor decision on both evidence and cost.
Keep exploring
- Can you use NLP on yourself? — the same sceptical lens applied to solo practice, where the placebo question is sharpest.
- 12 sentences you say about yourself that have no evidence — our most recent piece, and a practical companion: evidence-checking applied to your own beliefs rather than to the field.
Carolin Mallmann — Licensed Trainer of The Society of NLP, trained by Richard Bandler. She has trained 300+ people and runs English-language NLP trainings in Lisbon in groups of no more than ten.
Judge it for yourself: the next 7-day NLP Practitioner training in Lisbon
