Most people in this field are operating with a map full of wrong turns, and they do not know it. They repeat claims that failed testing forty years ago, promise outcomes nobody can deliver, and get dismantled by the first informed sceptic in the room. The practitioner who actually knows what replicates, and can say plainly what does not, is playing a different game against unprepared opposition. That is the unfair advantage, and it is available to anyone willing to do something almost nobody in this field does: read the evidence honestly.
Because make no mistake, the question underneath is real. "Is NLP scientific?" is the first question anyone should ask before paying for a training. It is also the question the NLP industry has spent decades dodging. The debunkers have their answer, and the seminar circuit has theirs. This post is the version we teach at the Institute: the honest one, in plain terms, because the honest version turns out to be the confident one.
The short answer
NLP is not a validated clinical treatment, and the specific testable claims it made in the 1970s and 1980s mostly failed testing. At the same time, a small body of research, including at least one meta-analysis, reports positive results for NLP-based coaching and therapy. Both statements are true, and a discipline that wants to be taken seriously has to hold both at once. If a trainer tells you either half, they are selling you something.
Here is why the honest position is the advantage and not the handicap. Every claim you cannot defend is a liability someone else can trigger. Strip those out, keep only what survives, and two things happen: you stop fearing examination, and your working set, smaller than the competition's, is entirely load-bearing. The practitioner who knows exactly why eye-reading failed can also tell you precisely what attention skills do work and demonstrate them. Opponents cannot attack what you have already conceded, and what remains conceded is, inconveniently for them, real.
What got tested, and failed
The early NLP founders made specific, testable claims. Two of them were studied repeatedly through the 1980s, and neither survived.
Eye accessing cues. The idea that the direction someone's eyes move reveals whether they are thinking visually, auditorily, or in feelings. Researchers tested this prediction over and over and found no reliable link between gaze direction and the kind of thinking going on. It does not work as a lie detector either, and any course selling it as one is selling a claim that failed testing. We cover the details in What NLP Cannot Do.
Preferred representational systems. The claim that each person has one dominant sense for thinking, and that matching it makes you persuasive. Repeated studies found people are far more flexible than the model assumed. The "one dominant sense" version did not hold up.
The most cited research on this came from Christopher Sharpley, an Australian psychologist, who reviewed the published NLP research in 1984 and again in 1987. His conclusion, published in the Journal of Counseling Psychology, was blunt: little empirical support for NLP's core claims, and what existed was mostly weak and practitioner-produced. If you have read "NLP has no evidence," you have read a summary of Sharpley.
What the research does show
It is not all one-sided, and honesty cuts both ways.
The 2015 meta-analysis. A meta-analysis titled "Evidence-based Neuro Linguistic Psychotherapy," published by Stipancic and colleagues in 2015, pooled studies of NLP-based psychotherapy and reported large effects for it. That sounds like vindication, so read the fine print: the number of studies was small, several were weak or unpublished, and the authors were inside the NLP community. It is real, and it is not the strong evidence the field pretends it is. It moves NLP from "no evidence at all" to "promising but under-tested." That is a meaningful difference, and we state it plainly.
Adjacent evidence. Much of what NLP actually does overlaps with better-researched practices. Reframing resembles cognitive restructuring in CBT. Well-formed outcomes resemble the goal-setting research in coaching psychology. The Ericksonian hypnosis that NLP borrowed has a real research base, mixed as it is. NLP did not invent these mechanisms; it packaged them in a memorable vocabulary. That is not a criticism, it is context.
The "35 years of research" question
Every NLP brochure says the field is "backed by 35+ years of research." The honest translation: NLP has had 35 years of marketing. Most of the research produced inside the community was never peer-reviewed, a large share was never published at all, and the published core is a few dozen studies of uneven quality. When someone quotes "35 years of research" at you, ask which studies, which journals, and which independent replications. Watch what happens.
Why the field overclaimed
The overclaiming is not an accident. NLP was born in the seminar boom of the 1970s, grew rich on the self-help circuit, and discovered early that "cure your phobia in ten minutes" sells tickets while "training attention and language patterns" does not. So the claims escalated: mind control, instant therapy, reading people like books. The honest practitioners got drowned out by the loud ones. That is why the reputation today is worse than the reality, and why a field with a thin evidence base is so easy to attack.
Where NLP genuinely helps
The evidence-supported uses are practical and modest:
- Communication and coaching. Rapport, questioning, reframing and state management are teachable skills with real, observable effects in conversations.
- Skill acquisition. Modelling, the actual spine of NLP, is a repeatable method for learning what excellent people do. That is what the founders originally did with Perls, Satir and Erickson.
- Self-change. For habits, procrastination, public speaking nerves and similar non-clinical problems, the patterns are genuinely useful to many people. The evidence here is anecdotal and mixed, and we say so.
- Not as treatment. For diagnosed mental health conditions, NLP is not a substitute for therapy with a licensed professional. That line is not negotiable.
How IONI handles the evidence
We teach the failures by name. Eye accessing cues, fixed representational systems, the mind-control fantasies: they are in the curriculum, with their evidence status attached, because a discipline that cannot name its own failures is a belief system. We teach the meta-analysis with its caveats, not as a trophy. And we train the skills that do transfer, with supervised practice and feedback over 24 weeks, not a seven-day certificate. That is our answer to the evidence problem: not louder claims, but better training. The ethics that keep the advantage clean are written out here, and the full list of what the field cannot do is its own piece.
If you want the full picture of what NLP is before you judge it, start with our plain-language guide to NLP. If the honest-evidence approach sounds like your kind of training, see the Spirit of NLP programme or begin with the free webinar.