Due to its inherent (and historical) preference for pragmatism over theory, its lack of formal and theoretical structure, and its lack of controls over usage, NLP doesn't always lend itself well to the scientific method. Equally (as scientific researchers have pointed out), attempts have also been greatly confounded by other factors including poor scientific appreciation of the NLP process being researched, unrealistic claims by some practitioners, and often a lack of high quality experimental design (such as failure to fully consider, control and understand all key variables).
This finding was supported in 1988, by both Heap and Druckman, who independently concluded that most studies to that date were "heavily flawed and that the "effectiveness of NLP therapy undertaken in authentic clinical contexts of trained practitioners has not yet been properly investigated.
NLP is not a science. It was not developed using the standard scientific method in the sense that its basic assumptions, models, and theories have, for the most part, never been tested scientifically. It is also comparable to heuristic problem-solving methods, in that the methods of NLP are tools for uncovering information and refining approaches, and the information uncovered is simultaneously information about the landscape as well as refinements to the heuristic algorithm which can help better identify what a solution might look like, and to find optimal 'next steps' to any of the many possible solutions.
However, this does not mean that NLP cannot be tested scientifically, especially as it relates to therapy. In this sense NLP can be considered to be a black box in that, while the internal workings may be mysterious, the quality of the output (results) may be evaluated statistically.
Actual clinical studies have been more productive, but many are merely suggestive or lack formal academic rigor. Equally (as researchers have pointed out), attempts have also been greatly confounded by many other factors including unrealistic claims by some practitioners, poor scientific understanding of the NLP process being researched, and often the lack of high quality experimental design (such as failure to fully consider, control and understand all key variables). Key issues expected or highlighted include:
Standard psychotherapy disciplines suffer from similar difficulties. In general, very well thought-out experimental design is necessary to test such subjects scientifically.
NLP practitioners test NLP in a practical environment much like a sports player tests a new skill in sports training - by integrating the skills and its concepts into their behavior, practicing, refining, and then observing the effect subjectively as to how it modifies their ability and competentce.
Skills in this context, for NLP, include skills of micro-observation of physiology, including the visible indicators of the body's autonomic systems, sensitivity to language patterns, awareness of the quality of rapport, and the like. Testing is performed within NLP on a small scale basis - observations are made, a decision is made what to try next, the results are observed. The T.O.T.E. model is one NLP formalization of this. NLP is also a generative system - that is, it contains the means to generate new extensions of itself. Thus a second way NLP practitioners test within NLP, is by their use of its approaches and philosophy to generate new shortcuts, possible approaches, or models describing how subjective awareness might be usefully considered, and then testing these in daily life to confirm if they seem to be useful.
In NLP, specific patterns (physical or mental) may be noted in a subject. Patterns of inner and external behavior are sought for their utility value. However, the belief that these - if found - are universal (for people, or even for a given individual) is not part of NLP.
As such, NLP practitioners do not so much test NLP itself, as much as they (subjectively) test their ability to better work with people using the approaches it suggests, and their satisfaction with NLP-based approaches in terms of their value to the task being undertaken.
It is important to recognize, that research -- both scientific and within NLP -- is susceptible to a variety of experimental errors. Readers should be aware of this if relying upon any given report, and confirm for themselves whether those concerned have taken adequate measures to control for known sources of error.
A large number of reputable bodies use NLP, including clinical, psychiatric, non-profit health, law enforcement, government, and education, giving rise to a significant number of sustained strongly positive reports.
A number of reports suggest that (also anecdotally) other users have encountered charlatans and low quality or charismatic trainers who place reliance upon emotional contagion rather than methodical formal practice. (See: History of NLP -- NLP buzz)
The Irish National Center for Guidance in Education's "Guidance Counsellor's Handbook sums up NLP's user evaluation, stating that:
However, in 1996, a US survey of clinical psychologists, measuring their opinions about different treatments, showed NLP to be one of the psychological therapies perceived by clinicians as being most pseudoscientific and questionable.
Sharpley (1984) undertook a literature review of 15 studies on the existence and effectiveness of preferred representational systems (PRS), at one time an important underlying principle of NLP, and found "little research evidence supporting its usefulness as an effective counseling tool" and no reproducible support for preferred representational systems and predicate matching. What neither the experimenters nor Sharpley realised was that they had totally misunderstood the NLP PRS model and that therefore the experiments were testing completely different claims than those made by NLPers. For example, many reviewers assumed that the NLP model was based on every person having a single, permnanent PSR; whereas in fact the NLP PRS model is based on the view that PRSs are entirely flexible and a person may shift from one to another as frequently as every 20-30 seconds or so. Thus Sharpley's findings, like those of Michael Heap's 1988 survey (based on some 70 studies), along with most of the experiments they reviewed, were more or less completely invalid.
Einspruch and Forman (1985) broadly agreed with Sharpley but disputed the conclusions, identifying a failure to address methodological errors in the research reviewed.They stated "NLP is far more complex than presumed by researchers, and thus, the data are not true evaluations of NLP" adding that NLP is difficult to test under the traditional counseling framework. Moreover the research lacked a necessary understanding of pattern recognition as part of advanced NLP training, there was inadequate control of context, an unfamiliarity with NLP as an approach to therapy, inadequate definitions of rapport and numerous logical mistakes in the research methodology.
Sharpley (1987) responded with a review of a further 7 studies on the same basic tenets (totalling 44 including those cited by Eispruch and Forman). This included Elich et al (1985) who tested the model that proposed a relationship between eye movements, spoken predicates, and internal imagery, and found no support for this model. They added "NLP has achieved something akin to cult status when it may be nothing more than a psychological fad" (p625)". Sharpley states that a number of NLP techniques are worthwhile or beneficial in counselling, citing predicate matching, mirroring clients behaviors, moving sensory modalities, reframing, anchoring and changing history, but that none of these techniques originated within NLP, saying "NLP may be seen as a partial compendium of rather than as an original contribution to counseling practice and, thereby, has a value distinct from the lack of research data supporting the underlying principles that Bandler and Grinder posited to present NLP as a new and magical theory". He concluded that as a counselling tool, the techniques and underlying theory unique to NLP, were both empirically unvalidated and unsupported but that "if NLP is presented as a theory-less set of procedures gathered from many approaches to counselling, then it may serve as a reference role for therapists who wish to supplement their counselling practice by what may be novel techniques to them."
A study by Buckner et al (1987), (after Sharpley), using trained NLP practitioners found support for the claim that specific eye movement patterns existed for visual and auditory (but not kinesthetic) components of thought, and that trained observers could reliably identify them. However, the study did not cover whether such patterns indicated a preferred representational system. They also made suggestions for further research. Krugman et al (1985) tested claims for a 'one-session' treatment of performance anxiety against another method and a control group and found no support for claims of a 'one-session' effective treatment.They argued for further research into NLP amongst other treatments that have "achieved popularity in the absence of data supporting their utility".
In 1988 a report by Druckman and Swets from the United States National Research Council, found that "individually, and as a group, these studies fail to provide an empirical base of support for NLP assumptions...or NLP effectiveness. The committee cannot recommend the employment of such an unvalidated technique". They also concluded influence techniques of NLP were unsupported (including matching representational systems to gain rapport). Moreover "instead of being grounded in contemporary, scientifically derived neurological theory, NLP is based on outdated metaphors of brain functioning and is laced with numerous factual errors". They conceded that the idea of modeling of expert performance "merits further consideration" but NLP itself was not included in a follow up study on modeling (amongst other matters) by Swets and Bjork (1991) except by way of acknowledgment for the idea which has been pursued through other disciplines.
Efran and Lukens (1990) stated that the "original interest in NLP turned to disillusionment after the research and now it is rarely even mentioned in psychotherapy".
Barry Beyerstein (1995) asserts that "though it claims neuroscience in its pedigree, NLP's outmoded view of the relationship between cognitive style and brain function ultimately boils down to crude analogies." With reference to all the 'neuromythologies' covered in his article, including NLP, he states "In the long run perhaps the heaviest cost extracted by neuromythologists is the one common to all pseudosciences—deterioration in the already low levels of scientific literacy and critical thinking in society. "
According to Von Bergen et al (1997) NLP was dropped from the experimental psychology research stream. They stated that "in relation to current understanding of neurology and perception, NLP is in error" and that "NLP does not stand up to scientific scrutiny" .
Carbonell and Figley of Florida State University Traumatology published an exploratory study (1999) on Visual/Kinesthetic Disassociation a component of NLP and three other novel treatments or power therapies for trauma (Thought Field Therapy or Eye Movement Desensitization and Reprocessing and Traumatic Incident Reduction), which was not designed to be a comparison study and the methodology used was the "systematic clinical demonstration (SCD) methodology. This methodology guides the examination, but does not test the effectiveness of clinical approaches". With reference to Brief Treatments for the Traumatized (including NLP) John Wilson states that while it is "adequately descriptive of the clinical procedures, there is little, if any, empirically validated dated outcome studies to substantiate a theory driven and research informed brief treatment (p. 173–207)."
Eisner (2000) in 'The Death of Psychotherapy', states that not "one iota of clinical research supports their (NLP proponents) claims. Apparently, no peer-reviewed researched has been published in over a decade. Moreover, there has been virtually no comparative research recently that assesses NLP's effectiveness." Eisner (2000) believes that with no clinical support, NLP proponents make grossly misleading claims about its effectiveness.
Evidence-based psychologist Lilienfield et al (2002)), describe NLP as "a scientifically unsubstantiated therapeutic method that purports to "program" brain functioning through a variety of techniques, including mirroring the postures and nonverbal behaviors of clients" and include it in their description "(Quick Fix + Pseudoscientific Gloss) x Credulous Public = High Income".
Devilly (2005) states that "at the time of its introduction, NLP was heralded as a breakthrough in therapy and advertisements for training workshops, videos and books began to appears in trade magazines. The workshops provided certification... However, controlled studies shed such a poor light on the practice, and those promoting the intervention made such extreme and changeable claims that researchers began to question the wisdom of researching the area further and even suggested that NLP was an untestable theory"..."NLP is no longer as prevalent as it was in the 1970s or 1980s, but is still practiced in small pockets of the human resource community. The science has come and gone, yet the belief still remains".
NLP is often associated with the work of the influential hypnotherapist Milton Erickson, upon whose techniques it was originally modelled to a large extent. However, other hypnotherapists have criticised the NLP interpretation of Erickson's work. Andre Weitzenhoffer, an influential Stanford researcher and former colleague of Erickson was an important critic of NLP. He rejects the NLP version of Ericksonian hypnosis, concluding that in terms of their evidence-base, “the neurolinguistic programming notions of Bandler and Grinder […] have very little substance and no empirical foundations.” (Weitzenhoffer, The Practice of Hypnotism, 2000: 108).
Albert Ellis, the founder of REBT, recently dismissed NLP as one of the “techniques that are avoided” in his approach, picking it out as his key example of a therapy rejected from evidence-based treatment because of its “dubious validity” (Dryden & Ellis, in Dobson, 2001: 331).
Michael Heap one of the UK's leading researchers on hypnotism, conducted a systematic review of the research literature on NLP and found, after analysis of over 60 research studies, that it was lacking in evidence,
Fortunately Heap, like Sharpley, was completely mistaken as to the true claims of the NLP PRS model, and consequently his comments quote about are completely invalid.
Examples of ongoing research include the following: an empirical study using a heuristic qualitative methodology in which the submodality change process was tested for treating grief and mourning and which suggested that manipulating certain submodalities can help the subject shift into more resourceful state and speed the healing process, a study incorporating a number of NLP behaviour change techniques (anchoring, isomorphic metaphor, goal setting) into a program for learning about and preventing the spread of AIDS which recommended that these tools be promoted and adopted internationally, and a review of several small studies on the effectiveness of Visual/Kinesthetic Dissociation (V/KD) which suggested that V/KD, although currently at an experimental level of efficacy and in need of further well-designed empirical study, may be a promising treatment for at least some forms of Posttraumatic Disorder.
NLP teaches that each learner is a unique individual with unique needs and backgrounds and encourages students take responsibility for their own states, and learning experience. According to Craft, "Neuro-linguistic Programming is best understood as a strategy which at first sight appears to draw on constructivist theories of learning. However, I have raised some problems in both the strategy itself (first), the lack of awareness of learning and performance styles, in that although it comes from the perspective that individuals respond uniquely to the world, it nevertheless offers an experiential approach to learning"
In response to Craft, NLP academics Tosey and Mathison (2003) expressed NLP in terms of the cybernetic epistemology of Gregory Bateson stating, "Early statements from the originators of NLP dismissed interest in articulating or acquiring theory, for example; `We have no idea about the "real" nature of things, and we're not particularly interested in what's "true". The function of modeling is to arrive at descriptions which are useful.' (Bandler and Grinder 1979 p.7). Bandler and Grinder's intent, perhaps, was to stay close to experience and avoid abstract discussion about truths of human experience. It seems that this stance has persisted, even if it is not shared by all leading NLP practitioners. We suggest that NLP may be regarded as a transdisciplinary (Gibbons et al 1994), in the sense that it draws on sources from academe and from elsewhere, and has been generated through application more than being deduced from axioms." They go on to state "It seems unarguable that to become regarded as academically credible there is a need for NLP to be more thoroughly theorized, particularly to consider how it relates to and differs from existing theoretical perspectives such as semiotics, phenomenology, discourse analysis, and more." Paul Tosey, University of Surrey, is the director of research project into the theory and application of NLP in teaching theory. Furthermore Mathison and Tosey (2002) add that this approach to learning and development appears similar in theory to Lev Vygotsky and constructivist learning theory and unrelated to computer programming or neuroscience.
In Whispering in the Wind Grinder & Bostic St Clair (2001) make suggestions about what needs to be done next to "improve the practice [of NLP] and take its rightful place as a scientifically based endeavor with its precise focus on one of the extremes of human bahavior: excellence and the high performers who actually do it." So at least one of the co-founders think that NLP can organise itself as a legitimate scientifically based discipline.
But note the difference between "a scientifically based endeavour" and an actual science. That is to say, between a discipline which can claim some supoorting evidence from science, and that which actually operates at a scientific level. the overall field of NLP, being a form of "pure" psychology, does not claim to be able to produce consistent results as are found in authentic science, primarily because it deals with human beings and therefore a host of uncontrolled and probably uncontrollable variables, will never be verifiable by use of the procedure known as the "scientific method."
"There appear to be some significant methodological problems with the existing research evidence, together with the dangers referred to of concentrating on specific techniques while taking insufficient account of context. These problems may hinge around the phenomenological nature of NLP and the attitude of its practitioners to research, as well as some of the inherent contradictions that arise from the theoretical bases on which NLP draws."