Does NLP Work? An Honest Look at the Evidence

Does NLP work? Partly, and not in the way it is usually sold. The communication skills at its core, building rapport, asking precise questions, changing your own state deliberately and setting a clear outcome, are ordinary and they do help most people who practise them. The specific claims that made NLP famous, eye movements revealing how someone is thinking, matching a person’s sensory “type”, a phobia removed in ten minutes, have been tested repeatedly and have not held up.

Neuro-Linguistic Programming is a model of communication and change built in the 1970s by Richard Bandler and John Grinder from close study of three unusually effective therapists. Fifty years on, it is taught worldwide and treated as pseudoscience by a good part of academic psychology. Both of those facts have reasons behind them. What follows is what the research actually shows, which parts survive it, and how to tell a serious NLP training from a sales pitch.

The claim	What the evidence shows
People have a preferred representational system you can detect and match to build rapport	Not supported. Christopher Sharpley’s 1987 review in the Journal of Counseling Psychology found no reliable support for the preferred system idea or for matching it
The direction of someone’s eye movements reveals whether they are picturing, hearing or feeling	Not supported. Repeatedly tested and repeatedly unconfirmed, including a 2012 study by Richard Wiseman and colleagues in PLOS ONE
NLP improves health outcomes such as anxiety, weight or pain	Insufficient evidence. A 2012 systematic review by Jaqui Sturt and colleagues in the British Journal of General Practice found little evidence for effectiveness
NLP as a whole is an evidence based therapy	No. Tomasz Witkowski’s 2010 analysis of the NLP research base found the majority of empirical results negative
Rapport, precise questioning, state work and clear outcomes improve communication	Plausible and consistent with wider psychology, though established outside NLP rather than by it

What does the research on NLP actually say?

Where NLP has been tested properly, the results are mostly negative, and across most of its claims it has barely been tested at all. Those are two different problems and they are often confused with each other.

The most cited review of outcomes is Sturt and colleagues in 2012, which searched for experimental studies of NLP applied to health outcomes, found ten that met the criteria, and concluded that there was little evidence that NLP interventions improve health related results. Tomasz Witkowski took a different route in 2010, analysing the NLP Research Data Base itself: of the articles with empirical results, the clear majority did not support NLP’s claims. Michael Heap reached a similar verdict on the core assertions as early as 1988.

Absence of research is not the same as proof that something fails. But NLP has existed since 1975 and has been commercially successful throughout, which is long enough that a thin evidence base is itself a piece of information. A field that was confident in its results would have produced them by now.

Which parts of NLP hold up best?

The parts that are not unique to NLP. Strip away the branded theory and what remains is a well organised collection of practices that mainstream psychology also arrived at, usually with better evidence attached.

1. Rapport and the working relationship. Across psychotherapy research, the quality of the alliance between two people is one of the more consistent predictors of outcome, largely independent of the method used. NLP’s attention to building that connection deliberately is pointed at something real.
2. Precise questioning. The Meta Model, which asks for the specifics that a vague sentence has dropped, closely resembles the Socratic questioning used in cognitive behavioural therapy, where it has a far stronger record.
3. State work and mental rehearsal. Deliberately recalling a resourceful state and rehearsing a performance in advance are standard sport psychology, with decades of evidence behind imagery and rehearsal.
4. Clear, well formed outcomes. Turning a vague wish into a specific, evidence based goal echoes the goal setting research of Edwin Locke and Gary Latham, which is among the more robust findings in organisational psychology.

This is a genuine criticism of NLP and also the reason a good training is still worth doing. The tools work; the theory wrapped around them is doing less than it claims. If you want the underlying model without the marketing, our pillar page on what NLP is and how it works lays it out plainly.

Which NLP claims fail, and why does that matter?

Three of them, and they fail badly enough to have shaped NLP’s reputation more than anything that works.

The first is eye accessing cues, the idea that where a person looks reveals which internal channel they are using or whether they are lying. It is NLP’s most recognisable claim and the one with the clearest negative record. Wiseman and colleagues tested the lying version directly in 2012 across three studies and found no relationship. Our glossary entry on eye accessing cues sets out what is left of the original observation.

The second is the leap from representational systems to fixed learner types. The observation that someone describes an experience in pictures or in feelings became, largely through people selling training to schools, the visual, auditory and kinaesthetic learning styles model. Harold Pashler, Mark McDaniel, Doug Rohrer and Robert Bjork examined it in 2008 and found almost no studies designed to test the claim properly, and no support in the ones that existed.

The third is the promise of speed: a phobia gone in one session, a belief rebuilt in twenty minutes. Some people do report fast change, and fast change happens in other therapies too. Selling it as the standard result sets up most participants to conclude that they are the problem when their own change takes months.

Why it matters is practical rather than academic. Overclaiming is the reason practitioners with real communication skills get dismissed unheard, and the reason a serious question about NLP usually gets a defensive answer instead of an honest one.

Why do so many people say NLP worked for them?

Because several ordinary mechanisms produce real change, and none of them require the branded theory to be correct. The change people report is often genuine. The explanation attached to it is the part that is probably wrong.

• Expectancy. People arrive at a training expecting change, having paid for it and taken time off. Expectation alone moves outcomes, which is why clinical trials go to such lengths to control for it.
• Sustained attention. Seven days of structured attention on your own patterns, with other people asking you direct questions, is unusual. Most participants have never had it before.
• Timing. People book personal development at their worst moments, and worst moments tend to be followed by improvement regardless of what happens next.
• Actual skill. Someone who practises asking better questions for a week does ask better questions afterwards. That is not a placebo. It is learning.

Honest trainers hold both halves of this at once: the results in the room are frequently real, and the story told about why they happened is often overstated. If you want a closer look at the evidence question specifically, the piece on whether NLP is scientifically proven goes through the studies one by one.

How do you tell a useful NLP training from a sales pitch?

By what the trainer says about the limits. The single most reliable signal is whether they will name what NLP does not do without being pushed. Beyond that, a short checklist covers most of it.

• Cure language for clinical conditions, or any guaranteed outcome, is a reason to walk away.
• Certification without supervised live practice is paperwork. You cannot learn to calibrate another person from a video.
• Ask how big the group is. Correction while you practise is what you are paying for, and it does not survive a room of sixty.
• Ask who licenses the certificate and check the licensing body exists independently of the trainer.
• Ask what the trainer thinks of the research. Someone who has never read a critical review of their own field is not a reliable guide to it.

Being able to answer those questions honestly is the reason we cap groups at ten on the seven day Society of NLP Practitioner certification we run in English in Lisbon, where the time goes on practice under correction rather than on demonstrations from the front.

One boundary is not negotiable. If you are dealing with anxiety, depression, trauma or any other clinical condition, NLP complements professional care and does not replace it. A Practitioner certificate is not a clinical qualification, and no reputable trainer will suggest that it is.

Carolin Mallmann, Licensed Trainer of The Society of NLP, Lisbon

Frequently asked questions

Keep exploring

• Is NLP scientifically proven?, for the individual studies behind the summary above.
• NLP for parents, for what the communication half of NLP looks like when it is applied at home.

---