Why Researching NLP Based Methodologies Can Be Problematic

In the last blog I introduced you to the history of the MAM 3 step rewind technique (MAM). We saw that it has its’ roots in the model of NLP, we explored the creators of the model and how the model itself now crops up in many different contexts; Therapy, Business, Education and the Entertainment Industry i.e Acting and performance (to name a few).

For a long time, research regarding the methodology has been seriously flawed. In the past, often the people seeking to conduct the research have not been skilled in the model itself, which has contributed to poor research design, leading to results that have had little value.

The early research conducted directly on NLP concepts was done in the 1980s and 1990s; little or no research has been done directly on NLP in the last decade or so (although that seems to be changing now). The majority of the studies that were done earlier addressed the concept of a primary representational systems (PRS), this is the theory that people are primarily visual, auditory, or kinesthetic and that knowing someones PRS can have impact on the depth of rapport experienced by the client. As the theory goes, if the therapist seeks to match the sensory descriptive words used by their clients’, rapport with the client will be enhanced significantly.

It probably needs saying at this point of our discussion that the original developers have in the past explicitly stated that NLP is fundamentally unverifiable by scientific research. One widely regarded leader in the field has even said that since NLP is about subjective experience, it is inherently untestable. This of course ignores the fact that dreams, the most subjective experiences that most of us will ever have, have been researched scientifically for decades. A variety of new methods of brain scanning make it now possible to do all sorts of experimental work on internal mental events, some of which are not even subjective experiences! For instance, brain scans have been used to detect when a decision is about to be made by a subject seven seconds before the subject becomes aware that they have made a decision.

As those of you with significant NLP training will already know, whether or not people have a PRS is not in any way central to the field of NLP. PRS does not really have anything to do with the effectiveness of the many methods that NLP practitioners have come to rely on to get results for people they work with who are wanting change in their lives.

What is this person’s PRS? This is just not a useful question to ask.

It is often more useful however to notice what sensory channel the client is using at the moment, or what sensory channel underlies the problem. For example it can be useful to notice that someone’s unpleasant feelings result from a critical inner voice, or to notice that many large and close movies of things that need doing leads to feelings of overwhelm.

Investigating Primary Representation System is a bit like Nasrudin looking for his lost car keys under the street lamp because the light is better here, even though he lost them somewhere else. PRS was perceived to be an easy thing to study, but the results of those studies do not tell us anything about the field of NLP.

To quote Steven Andreas when writing about this subject:

It’s also worth noting that the studies themselves were often full of research errors. The questionnaires used in an attempt to assess PRS often had confusing self-report questions like, “Do you see yourself as a feeling person,” or “Do you feel you are an auditory person?” As that kind of question clearly reveals, most experimenters were not trained in NLP, did not understand what they were researching, and did not use anyone trained in NLP as a consultant to review their experimental protocols. As a result, there was no controls of the language used in the studies, nor control of nonverbal confounding variables such as gestures or voice tone.

For instance, when matching a subject’s visual predicate with a sentence like, “I see what you mean,” a higher-pitched voice, looking up, or a pointing gesture in the upper visual field will be congruent with visual processing, and be more likely to result in rapport. However, a lower-pitched voice, looking down, or a palm-up gesture in the lower visual field will be incongruent, and be less likely to lead to rapport. (Visual processing is typically accompanied by a high voice tone, looking up, and pointing gestures, while kinesthetic processing is often accompanied by a lower voice tone, looking down, and palm-up gesturing.)

As a result of these kinds of mistakes, most of the research was very poor quality. Not surprisingly, there is very little direct academic experimental support for NLP. A research committee working for the United States National Research Council in 1988 found little if any evidence to support NLP’s assumptions or to indicate that it was effective as a strategy for social influence. “It [NLP] assumes that by tracking another’s eye movements and language, an NLP trainer can shape the person’s thoughts, feelings, and opinions. There is no scientific support for these assumptions.”

To summarise then, the research that has been done in the past was in the main asking the wrong questions; the researchers asking the questions did not understand what they were trying to measure. They frequently ignored linguistic and behavioural confounding variables, so it comes as no surprise that the results were often negative or inconclusive.

As someone who has practiced and studied the model of NLP for over 10 years, I find this debate regarding the model and research very frustrating. The reason for my frustration is that in my experience all of the NLP core processes include specific testable outcomes, detailed systematic protocols for different kinds of problems, and clear operational tests in sensory-based experience to determine when a client has reached their outcomes. In addition, many NLP processes can be completed in a single session of an hour or less.

Because of this it seems to me that NLP would be much easier to research than most therapies which are much less structured and usually take place during many sessions over a period of weeks or months. Scientific research definitely needs to be done (and is now being done thankfully) in order to confirm or disconfirm the various processes and understandings that are typically included in the term NLP. Often in our trainings Mia and I are asked about research. The question might be framed like this: Why isn’t the MAM 3 step rewind technique in the NICE guidelines for working with PTSD?

The conversation we have had so far goes some way to explain why. To go further, it is interesting to note, I think, that cognitive behavioural therapy (CBT) does find a place in the NICE guidelines. As a methodology of therapy it has some similarities to NLP. It tends to focus entirely on the content of the clients’ internal auditory dialogue the words that people say to themselves.

Yet, CBT ignores distinctions like the volume of the internal voice, the location of the voice in personal space, the direction from which the voice comes into personal space, its tonality and tempo etc. Often changing these process parameters can have a much greater impact on a clients’ subjective experience than just changing the content alone.

In my next blog I will start to explore the research that has been done more recently, in the UK and in the USA.


  • Angela says:

    Interesting, the mind is fascinating how it affects our every living cell. Having not come from an NLP background l’m keen to explore the concepts. Looking forward to the next blog.

  • >