NLP is controversial for three reasons that usually get argued as one. Its research record is thin and, where studies exist, largely unsupportive. Several of its early signature claims — most famously that eye movements reveal how someone is thinking — have failed replication. And the field has no single licensing authority, so the quality of what is sold under the name “NLP” ranges from serious training to a weekend certificate bought online.
Neuro-Linguistic Programming is a model of communication and change assembled in the 1970s by Richard Bandler and John Grinder from observing therapists they considered exceptional. It is best described as a practical toolkit with weak scientific credentials — not a validated therapy, and not a fraud either. Most of the heat in the argument comes from people defending or attacking a claim NLP never needed to make.
| The criticism | Where it actually lands |
|---|---|
| “There is no evidence NLP works” | Largely fair. Reviews since the 1980s found little support for its testable claims |
| “The eye movement model is pseudoscience” | Fair. It has repeatedly failed testing, most publicly in 2012 |
| “NLP is manipulation” | Overstated. Influence skills can be misused; so can any communication training |
| “Anyone can call themselves an NLP trainer” | Entirely true. The field is unregulated and certification quality varies wildly |
| “It was oversold” | True, and self-inflicted. Cure-in-ten-minutes marketing did lasting damage |
| “Serious psychologists dismiss it” | Many do. Some of the underlying tools appear in mainstream practice under other names |
What does the research actually say?
The most-cited assessments are unfavourable, and there are three worth naming rather than gesturing at. In 1988 the US National Research Council, asked by the Army to evaluate techniques for improving human performance, examined NLP and reported no significant evidence for the influential aspects of the model (Druckman and Swets, Enhancing Human Performance). Christopher Sharpley’s reviews in the Journal of Counseling Psychology (1984 and 1987) found the preferred representational system hypothesis unsupported. Tomasz Witkowski’s 2010 review in the Polish Psychological Bulletin surveyed the empirical literature and concluded that the majority of studies did not support NLP’s claims.
Two qualifications matter, and neither rescues the field. First, most of that research tested a narrow set of laboratory-friendly claims — eye movements, matching someone’s sensory vocabulary — rather than the composite thing that happens across seven days of practice. Second, that is precisely the problem: after fifty years, a discipline that wanted to be taken seriously would have funded the harder studies. It mostly funded marketing instead. A closer look at what the evidence for NLP does and does not show is the fair-minded version of this argument.
Which NLP claims have genuinely failed testing?
Two, clearly. The eye accessing model — the idea that looking up-left versus up-right reliably indicates remembering versus constructing, and therefore truth versus lying — was tested directly by Richard Wiseman and colleagues in PLOS ONE in 2012 and found no supporting relationship. The preferred representational system claim, that matching someone’s “visual” or “auditory” language improves rapport and outcomes, failed repeatedly in Sharpley’s reviews.
Honest practitioners teach both as history. In our trainings, eye accessing cues are presented as an observation from the 1970s that did not survive testing, useful mainly as a lesson in how a plausible pattern becomes a doctrine. If a trainer still sells eye movements as a lie detector, that tells you more about the trainer than about NLP. — Carolin Mallmann, Licensed Trainer of The Society of NLP, Lisbon
Is the manipulation accusation fair?
Partly, and not in the way most critics mean. NLP teaches rapport, pacing and language patterns that increase influence, and influence skills are neutral in the same way a sharp knife is neutral. The accusation lands when trainings teach covert technique with no discussion of consent or ecology — and some do. It misses when it assumes that noticing how someone speaks and adjusting your own delivery is inherently sinister; that is what good listeners have always done.
The reasonable test is whether the person would object if they saw the transcript. Where influence ends and manipulation begins is the part of the debate worth having, and it is an ethics question rather than an evidence question.
Why does the certification free-for-all make it worse?
Because it removes any floor. NLP is not a protected title in any country we know of, so a two-hour video course and a seven-day assessed training can both issue something called a Practitioner certificate. Prospective clients cannot tell them apart, critics reasonably judge the field by its worst examples, and the good trainings carry the reputational cost of the bad ones.
- Who certifies, and did they exist before the course? The Society of NLP, founded by Richard Bandler, is one of the original bodies. Many “institutes” are one person and a logo.
- How many live, assessed practice hours? If the answer is zero, the certificate documents attendance, not skill.
- How large is the group? Practice quality collapses when one trainer supervises forty people.
- What does the trainer say NLP cannot do? An answer that comes easily is a good sign. No answer is a warning.
- Is there a claim to cure diagnosed conditions? If yes, walk away — that is both an evidence problem and a safeguarding problem.
None of this makes the criticism go away, and it should not. It does mean the sensible response to “NLP is controversial” is to ask which NLP, taught by whom, for how long. Our own answer is a seven-day certification week in Lisbon with a maximum of ten participants, which is a format choice, not a defence of the field’s evidence base.
So what survives the criticism?
The parts that were never exotic claims to begin with. Setting outcomes in specific, sensory terms. Asking precise questions about vague statements. Noticing what changes in someone’s face and voice while they talk. Deliberately shifting your own state before a hard conversation. These overlap heavily with methods that mainstream psychology has independently validated under other labels — goal specificity, Socratic questioning, exposure, self-regulation.
What that leaves is a repackaging argument rather than a magic argument: NLP assembled a usable, teachable kit, gave it idiosyncratic names, and then oversold it. If you want the plain version of what the kit contains, our explanation of what NLP is and how it works avoids the claims that got the field into trouble.
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 trauma, depression, an anxiety disorder or any diagnosed condition, work with a licensed clinician.
Frequently asked questions
Is NLP pseudoscience?
Several specific NLP claims meet the definition — they were presented as established, and they failed testing without being withdrawn. The label fits the claims better than it fits the whole toolkit, parts of which are ordinary communication practice with unusual names. The honest position is that NLP is not a science and has not behaved like one.
Is NLP recognised as a therapy?
No. It is not a regulated psychotherapy in any major jurisdiction, and it does not appear on evidence-based treatment lists such as those maintained by national health bodies. NLP practitioners are not clinicians, and a responsible one will say so and refer on when a client needs clinical care.
If the evidence is weak, why do so many coaches still use it?
Because the tools are teachable in days and produce visible movement in ordinary conversations, which is a low bar for a coach and a high bar for a laboratory. That is a real observation and a weak defence: plenty of ineffective methods also feel like they work. It explains the persistence without settling the argument.
Does the controversy mean NLP does not work at all?
No. It means the specific mechanisms NLP proposed are unproven and some are disproven, while the practices built on them have not been properly tested either way. Treat claims about outcomes as unverified rather than as established fact, and judge a training by what it teaches you to do rather than by what it promises.
Should I avoid NLP training because of the criticism?
Not automatically — but go in with the criticism in hand. Choose a training that names its certifying body, gives you many hours of supervised live practice, and is willing to state the limits of the method. Avoid anything promising rapid cures, guaranteed income, or techniques that work on people without their knowledge.
Who are NLP’s best-known critics?
Christopher Sharpley, whose reviews dismantled the representational systems claim in the 1980s; Tomasz Witkowski, whose 2010 review classed NLP as largely unsupported; and Richard Wiseman, whose 2012 study tested the eye movement model and found nothing. Reading them directly is more useful than reading either NLP marketing or online summaries of the debate.
Keep exploring
- Is NLP scientifically proven? — the evidence question on its own, study by study, without the ethics and marketing arguments tangled into it.
- Is NLP manipulation? — the other half of the controversy: where influence stops being fair, and the consent test worth applying to any technique you learn.
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.
See how the method is actually taught: the 7-day NLP Practitioner certification in Lisbon
