Selected Alternative Training Techniques in HRD

  1. W. Von Bergen, Barlow Soper, Gary T. Rosenthal, Lamar V. Wilkinson

Over the years a number of training techniques and procedures have been developed that are not part of the mainstream but are believed by some to have utility for organizations trying to enhance human performance. This article discusses four of these alternative techniques–subliminal self-help, mental imagery and practice, meditation, and Neurolinguistic Programming (NLP)–and examines the contributions of each from a scientific perspective. With the exception of mental practice, there is a paucity of data to demonstrate convincingly whether these alternative techniques promote or enhance individual or organizational effectiveness. It is concluded that effective professional practice depends on scientifically derived research results.

Training and development has historically been plagued by charlatans hawking this or that program or approach as having a significant positive impact on profits, effectiveness, absenteeism, turnover, and so on. Fads and techniques that come and go abound in this discipline: participation training is the style for a few years, then total quality management training surfaces; management by objectives comes and zero defects goes; T-groups are the darling of training professionals for a while; time management is “in” and job enrichment is “out”; Kepnor-Tregoe is popular for a time and then the one-minute manager takes its turn in the spotlight. Although these approaches are generally considered mainstream, others that have app eared are less traditional but also have significant followings proclaiming their effectiveness.

As a class, the latter techniques are considered extraordinary; they are developed outside of mainstream science and are accompanied by extreme claims for high effectiveness (Austin and Miller, 1992). These claims are often based on a handful of testimonials or individual cases instead of controlled experimentation and peer review of results.

An analogous situation has developed in medicine as a wide range of techniques known collectively as alternative medicine have gained increased emphasis. Some of these techniques have been around for hundreds, even thousands, of years; others, like biofeedback for example, are modern techniques that have filled useful niches alongside orthodox medicine. It is estimated that a third of American adults spent some $13.7 billion in 1996 on a bewildering array of such techniques as acupuncture, colonic irrigation, homeopathy, hypnotherapy, aromatherapy, Shiatsu massage, music therapy, naturopathy, herbal medicine, and therapeutic applications of electromagnetic fields (Langone, 1996). In 1996 the National Institutes of Health (NIH) classified alternative medicine as an unrelated group of nonorthodox therapeutic practices that often have explanatory7 systems that do not follow conventional biomedical explanations. NIH even created an Office of Alternative Medicine to facilitate the evaluation of alternative medical treatment modalities in order to determine their effectiveness, educate the conventional scientific biomedical community in alternative medical treatment, and help integrate treatments into mainstream medical practice.

Although no governmental office has been established to investigate alternative performance enhancement practices in business and industry, this paper reviews the validity and potential effectiveness of several alternative training and development performance enhancement techniques. The list is not meant to be exhaustive, but it represents a sample of currently popular techniques gleaned from the management training and development literature (Druckman and Bjork, 1991). These techniques were deemed worthy of investigation by the National Research Council (Druckman and Bjork, 1991; Druckman and Swets, 1988) and were described as offenng “the potential to accelerate learning, improve motor skills, alter mental states, reduce stress” (Druckman and Bjork, 1991, p. vii), among other positive claims. Following NIHs definition of alternative medical practices, we classify these alternative training and development techniques as an unrelated group of nonorthodox, individual or organizational performance enhancement practices, often with explanatory systems that do not follow conventional scientific explanations. Literally dozens of such practices could be reviewed.

The current article critically examines four popular techniques of this nature: subliminal self-help, mental practice, meditation, and Neurolinguistic Programming. As for subliminal self-help, it has been estimated that over two thousand companies produce subliminal materials and that they generate over $50 million in annual sales (“Sound,” 1990). Regarding the use of mental practice, Dnskell, Copper, and Moran (1994) and Whetstone (1995) have noted that, though popularized in sports, mental practice is increasingly favored by business and other organizations. It has been stated that numerous organizations use meditation, including USA Global Link and Telegroup, Inc. (Schmidt-Wilk, Alexander, and Swanson, 1996). And, in their book, O’Connor and Seymour (1990) list over eighty training organizations in the United States using Neurolinguistic Programming. Together, the four techniques represent some of the more popular alternative approaches to performance enhancement. It is not unusual to talk with someone who uses one of them or knows of someone or some business that does.

The purveyors of these methods often make extreme claims for their techniques, yet generally do not document controlled results to justify such claims. According to Rosen (1993), psychological self-help is big business, and many practitioners rush to market with exaggerated product claims. Sometimes they rationalize their actions by saying they “know” their interventions work because they have seen it.

Finally, these four were selected for study because they have been widely touted and sold to business, industry, and government (Druckman and Bjork, 1991) and have received considerable attention in the popular press.

Subliminal Self-Help Products

Build self-confidence! Lose weight! Reduce pain! Quit smoking! Speak effectively! Read faster! Improve memory! These are but a few of the subliminal programs available to assist individuals in improving some aspect of their lives. According to Oldenburg (1990), about two thousand individuals or companies in the United States and Canada produce subliminal self-help products with retail sales estimated at $50 million annually. The subliminal industry is significant and, by most accounts, getting bigger (McGarvey, 1989; Natale, 1988; Oldenburg, 1990).

Subliminal is defined as below the threshold of consciousness. Typically what happens is that messages are hidden visually–presented for about one- sixtieth of a second–in a videotape (Smith, 1993) or auditorily–average sig- nal-to-noise ratios of -10 to – 20db–in an audiotape (Bordon and Harris, 1984). The messages suggest improvement in a selected area. The premise underlying subliminal stimulation is that a person can effortlessly accomplish in a short time what others struggle but fail to do in a lifetime, hence making these mass-marketed tapes very attractive. For example, Gateways Institute’s audiotapes advertise: “Subliminal tapes work, so you don’t have to. . . . Simply play the tapes while you work, play, drive, read, exercise, relax, watch TV, or even as you sleep. No concentration is required for the tapes to be effective. They work whether you pay attention to them or not” (Druckman and Bjork, 1991, pp. 107-8).

Audiotapes appear to be the most popular subliminal approach (Merikle and Skanes, 1992). Most of the commercially available subliminal audiotapes have a similar format. When the tapes are played, all that a listener consciously perceives are background sounds, such as music or ocean waves with the occasional cricket or seagull. Although each company’s tapes contain a unique mix of background sounds, different tapes produced by the same company are often indistinguishable. Presumably, what separates the many different tapes produced by each company are the embedded subliminal messages that are impossible for a listener to hear consciously in the context of the background.

Do these tapes work? In a word, no. A comprehensive review of the subliminal self-help literature conducted by the Committee on Techniques for the Enhancement of Human Performance of the National Research Council (Druckman and Bjork, 1991) unequivocally concluded that there is neither theoretical foundation nor experimental evidence to support claims that subliminal self-help tapes enhance human performance. Remarkably, a number of studies are very clear in demonstrating the ineffectiveness of subliminal tapes in modifying a wide range of behavior (Auday, Mellett, and Williams, 1991; Greenwald, Spangenberg, Pratkanis, and Eskenazi, 1991; Merikle and Skanes, 1992; Russell, Rowe, and Smouse, 1991).

These findings fly in the face of the positive testimonials given by individuals using the tapes, but there are several reasons to question the conviction that subliminal suggestions are in any way responsible for, or even play a part in, any self-perceived behavioral improvements. In the first place, by buying and using a subliminal self-help product, an individual demonstrates not only a desire for personal enhancement but also a commitment to change his or her ways. The very act of making such a commitment, of mobilizing oneself into action, may be therapeutic in its own right, in much the same way that some people realize marked improvements in behavior after they register for, but before they receive, psychotherapy (Rachman and Wilson, 1990). For these and other reasons testimonials are not accepted evidence by any science (Stanovich, 1996).

A second, related reason that subliminal tapes may seem to have beneficial effects has to do with the social psychological phenomenon of effort justification; that is, the harder we work at something, the more we like it (Penrod, 1983). After spending money to buy a subliminal tape and using it daily for several weeks many people would be reluctant to admit to themselves or to others that they had wasted their money and time. They would instead be motivated to detect any sort of change in some aspect of their lives in order to rationalize their purchase (Conway and Ross, 1984).

A third reason for attributing positive effects to subliminal tapes relates to expectancy effects, as clearly demonstrated by Pratkanis, Eskenazi, and Greenwald (1990; see also Greenwald, Spangenberg, Pratkanis, and Eskenazi, 1991). These researchers recruited individuals through a newspaper advertisement soliciting volunteers who were especially interested in subliminal self-help tapes and thus were likely to resemble the people most likely to purchase such products. On the first day of the study, the participants completed tests measuring their self-esteem and memory ability. They were then given a commer- dally available tape either for improving self-esteem or for enhancing memory. The intriguing aspect of this study was that only half of the individuals, selected randomly actually got the tape they thought they were getting while one-quarter received the memory enhancement tape mislabeled as the one to improve self-esteem and one-quarter received the self-esteem tape mislabeled memory enhancement.

The participants took their tapes home and listened to them for five weeks, as recommended by the manufacturer, and then returned to the experimenters who once again tested their self-esteem and memory. Although no differences were found–either positive or negative–on any measure of selfesteem or memory, many study participants believed and reported otherwise. Approximately half of those who thought they had received the self-esteem tape, regardless of whether they had actually received it, stated that their selfesteem had risen; similarly, about half of those who presumed correctly or not that they had received the memory tape asserted that their memory had improved as a result of listening to it. The title of the Pratkanis, Eskenazi, and Greenwald (1990) investigation–“What you expect is what you believe, but not necessarily what you get”–seems fitting.

In summary, a critical review of subliminal self-help training clearly demonstrates that this alternative technique to enhance human performance is not effective.

Mental Practice

A second alternative training and development technique is mental practice. Mental practice refers to cognitive rehearsal of a task in the absence of overt physical movement. When a musician practices a passage by thinking it through or when an athlete prepares for an event by visualizing the steps required to perform the task successfully, he or she is engaging in mental practice. Richardson provided the standard definition of mental practice as “the symbolic rehearsal of a physical activity in the absence of any gross muscular movements” (1967, p. 95). Such techniques also have been called imaginary practice (Perry, 1939), covert rehearsal (Corbin, 1967), symbolic rehearsal (Sack- ett, 1934), and introspective rehearsal or conceptualization (Egstrom, 1964).

In these studies, individuals are typically instructed to sit quietly, not move, and see themselves performing the task successfully from start to finish. Usually a control (no practice) group is included, as well as a group that actually practices the task physically. At a given period following the mental practice or physical practice sessions, actual performance is assessed. If the performance of the mental practice subjects exceeds that of the control subjects, mental practice is said to have a positive effect on enhancing performance.

Does mental practice really enhance performance? An excellent study by Driskell, Copper, and Moran (1994) addressed this question. These researchers conducted a comprehensive analysis of the literature on mental practice to determine the effect of mental practice on performance and identify conditions under which it is most effective. Results indicated that it has a positive moderate but significant effect on performance, and that its effectiveness is a function of the type of task, the retention interval between practice and performance, and the length or duration of the mental practice intervention.

More specifically, Driskell, Copper, and Moran found that mental practice is an effective means for enhancing performance. However, the data indicated that mental practice is less effective than overt physical practice. The authors suggested that this finding is not altogether surprising given what mental practice provides and what it does not. Mental practice offers the opportunity to rehearse behaviors and to code behaviors into easily remembered words and images to aid recall. Mental practice does not offer direct knowledge of results or visual and tactile feedback. Statistically, mental practice accounted for less than 7 percent of the variance in performance. Though this may not sound like much, how many of us would be thrilled with an increase in profits of nearly 7 percent? It might mean the difference between keeping a job or being asked to move on.

Second, Driskell, Copper, and Moran found that type of task is a significant moderator of the effectiveness of mental practice. Not surprisingly, the effect of mental practice on performance is stronger the more the task involves cognitive elements (for example, observing, reading, problem solving) as opposed to physical elements (exerting force, applying speed and power, maintaining balance). But even though mental practice is better suited to tasks that are cognitively laden, it can produce significant effects in a wide range of tasks, from determining volumetric analyses of chemical substances (Beasley, 1978) to welding (Hackler, 1971).

Third, the positive effect of mental practice on performance declines over time. Driskell, Copper, and Moran found that after approximately two weeks the beneficial effects of mental practice are reduced to one-half of their original magnitude, and after approximately three weeks the increase in performance due to mental practice substantially dissipates. These estimates provide practical guidelines for implementing mental practice: to gain maximum benefits of mental practice, one should implement refresher training on at least a one- to two-week schedule.

Fourth, these researchers found that expenenced people benefitted equally well from mental practice, regardless of task type. Novice subjects benefitted more from mental practice on cognitive tasks than on physical tasks. This result is consistent with Ryan and Simons’ (1983) argument that if an experienced individual has already learned the component motor skills of a physical task then mental practice may be sufficient to enhance performance without additional physical practice and feedback. But for novices, who have not formed an approximation of the skill, the symbolic rehearsal provided by mental practice

may not be sufficient to guide performance. This suggests that mental practice may be more effective, everything else held constant, if novice subjects are given schematic knowledge before mental practice of a physical task.

Finally, Driskell, Copper, and Moran (1994) established that more mental practice is not necessarily better. They suggested that approximately twenty minutes total duration may be an optimal mental practice intervention. This is consistent with many researchers who have held that extended mental practice may lead to a loss of concentration and that there is an optimal length for such interventions. For example, Corbin (1972, p. 106) argued that “relatively short” practice sessions are optimal.

In summary, it appears that mental practice is a moderately effective means for enhancing performance in some circumstances, though it is less effective than physical practice for some people. Thus, for dangerous tasks, tasks in which there are seldom opportunities for physical practice, or as a means of supplementing normal training, mental practice should be considered as a training alternative.


A third technique is meditation. Although for many people meditation conjures up visions of magic carpets and long-haired yogis sitting on mountain- tops, it has been scientifically defined as a class of techniques designed to influence an individual’s consciousness through the regulation of attention (Druckman and Bjork, 1991). Meditation typically requires that people lie quietly or sit in a particular position; attend to their breathing; adopt a passive attitude; be at ease; and, frequently, repeat aloud or to themselves a monosyllabic word, phrase, or sound, referred to as a mantra, typically for twenty minutes a day, twice a day, morning and evening.

Usually, meditation is viewed as a relaxation technique that can be used to cope with existing stress or prevent stress from occurring. But it can function in other ways. The popularity of meditation is a relatively recent phenomenon in the United States, although forms of meditation have been practiced in many Eastern countries for centuries. Meditation is an altered state of consciousness in which the individual may try to increase awareness and achieve bodily relaxation as well as relaxation of the mind (Benjamin, Hopkins, and Nation, 1994). Most types of meditation can be placed in one of two classes. Opening-up meditation requires the subject to achieve a clear mind by eliminating all thoughts and thus becoming especially receptive to new experiences. In concentrative meditation, the person focuses attention intensively on a single sound, idea, object, or action of the body, such as breathing (Naranjo and Ornstein, 1977).

Meditation originated in religious settings. Meditation and contemplation in the Jewish tradition allowed the concentration necessary for reading and

interpreting the Torah. This practice also runs through Indian religious history as far back as the third millennium b.c. (Bloomfield and Kory, 1976). One meditation technique, Transcendental Meditation (TM), was introduced to the United States in 1959 by the Indian teacher Maharishi Mahesh Yogi. It is by far the best known and most often practiced of the various approaches to meditation. Maharishi chose to use the term “transcendental,” meaning “going beyond,” to signify that TM takes one beyond normal wakeful experience to a state of restfulness that is also characterized by a heightened sense of alertness (Adeste, 1979).

TM creates an opportunity to disengage briefly from the unending cascade of thoughts, emotions, sensations, and perceptions. The meditator experiences quiet levels of the mind while becoming increasingly aware of the unbounded nature of this awareness in the absence of objects. This rest offers the opportunity for a variety of spontaneous and regenerative changes throughout the entire nervous system (Denniston and McWilliams, 1975).

A number of claims have been made for the benefits of meditation, including elimination of depression and anxiety, greater resistance to fatigue, improved sexual potency, greater creativity, stronger social skills, and better mental efficiency. More relevant to business, meditation advocates (for example, Gordon, 1991; К or)’, 1976; Schmidt-Wilk, Alexander, and Swanson, 1996) claim that executives and employees work together more effectively and accomplish greater productivity after commencing meditation, that productivity and sales dollars per employee almost double, and that some organizations have indicated that meditation programs generated so much enthusiasm that additional courses were opened to all executives and employees. Nice to say, wonderful to hear, but such assertions and testimonials are not scientifically rigorous and do not warrant confidence.

The problem, from a scientific perspective, concerns proper experimental controls. Controls for distraction or for just sitting or lying quietly and undisturbed are seldom found in the studies that demonstrate such positive benefits as reduction in respiration and heart rate and in experience of stress, anxiety, or anger (Cheaper and Giber, 1978; Delmonte, 1985). Indeed, when a resting-only control group is present, there is no evidence that reductions in heart rate, respiration rate, skin conductance fluctuations, or blood pressure are any less for that group than for experienced meditators (Holmes, 1984). In other words, resting in bed for twenty minutes produces the same reduction in somatic arousal as engaging in twenty minutes of meditation. There are also some reports of benefits in blood pressure reduction for borderline hypertensives (for example, Benson, 1975; Patel, 1973), but in most studies combined use of other techniques, such as relaxation, precludes clear attribution of any positive effects to meditation alone (Delmonte, 1984, as cited in Brener and Connally, 1986).

As for the highly publicized feats of some yogis who can remain buried for many hours without suffocating, for example, these are probably due not to any special properties of meditation but rather to these individuals’ confidence in their ability to slow down their respiration rate as well as their faith that they can survive the ordeal if only they do not panic (Druckman and Bjork, 1991). Or the reports may simply be fraudulent, like the repeated claims of yogi levitation (Stein, 1989).

Finally, two reviews of meditation research (Brener and Connally, 1986; Druckman and Bjork, 1991) found that in scientifically controlled studies meditation does not reduce somatic arousal any more than does simply resting quietly. In summary then, it must be concluded that there is no scientific evidence to support the unique benefits attributed to meditation by its promoters, who appear to be more concerned with demonstrations of the experience than with experimental evaluations of the procedure.

Neurolinguistic Programming

“Double your chances of success: A powerful new psychological technique called NLP can help us get what we really want–and need–out of life” (Robert Harding Syndications, 1995, p. 69).

“Cure phobias and other unpleasant feeling responses in less than an hour . . . cure many physical problems–not only most of those recognized as ‘psychosomatic’” (Stevens, 1979, p. ii).

“The most powerful and effective technology in human communication and change» (brochure, Jacobson, 1995, no page numbers).

Extraordinary claims, to be sure, but these are but a few of those being espoused for Neurolinguistic Programming (NLP), the final alternative training and development technique we will review in this article. NLP is a system of procedures that purports to enable people to increase their effectiveness in communicating with and influencing others. It was developed in the 1970s by Bandler and Grinder (for example, 1975; also Bandler and Grinder, 1979; Grinder and Bandler, 1976). Training in NLP is usually provided via workshop format; the third quote above is actually from a brochure advertising one such training opportunity.

NLP was derived through observations of three psychotherapists who Bandler and Grinder judged to be particularly effective and was formulated as a means of understanding and facilitating similarly high levels of interaction in others. NLP is founded on the idea that people communicate from a limited number of sensory-oriented representational systems. A representational system is a person’s typical, usual, and preferred way of interacting with the world. Some people are visually oriented, others are sound oriented, and still others are feeling oriented. When a person uses language that ties into the representational system used by another, that person is supposed to feel more understood and, ultimately, to be more easily influenced. Examples include statements like “I see what you mean” when conversing with a visually oriented person, “It sounds to me like. . . .” when communicating with a hearing-oriented individual, and “I/eel we should. . . when speaking with a kinesthetically oriented person. Thus, each phrase uses a verb linked to one of the representational modalities. One can cue off the words and match the others system with one’s replies or monitor the direction of eye movement, which is the most noteworthy nonverbal indication of an individual’s preferred representational system.

Bandler and Grinder said that these eye movements are remarkably fixed and that certain directions consistently correspond to the three representational orientations, as well as to certain thought processes, information recall, or image construction. Through NLP training it is stated that individuals can gain proficiency at identifying representational modes and use both that knowledge and an understanding of thought processes to establish high levels of interpersonal competency via language that mirrors others’ modes, that is, using the same sensory terms. But more importantly, applying this information via specific techniques like “anchoring” and “reframing,” it is said that a person also can readily and to a significant degree influence or direct another person’s thoughts, feelings, and opinions.

Three questions should be asked of NLP Is it based on knowledge of how the neurological and perceptual systems work? Do the proposed relationships among NLP variables exist? And does NLP produce the results it claims? In relating NLP to current understandings of neurology and perception, Druck- man and Swets (1988) and others (Bertelsen, 1987) believe NLP to be in error. 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 (Druckman and Swets, 1988).

Bandler and Grinder stated that they were not interested in establishing scientific validation of NLP but instead intended to portray what works. Hence, the authors only present anecdotal and testimonial data to support their suppositions and the relationships among NLP variables and concepts. Even an elementary text on scientific method (for example, Stanovich, 1996) details the myriad pitfalls of such a methodology and descnbes its irrelevance to legitimate theory building. Where controlled studies have been performed attempting to test NLP hypotheses–like the proposed relationship between eye movement direction and representational system–they consistently have failed to do so (for example, Bliemeister, 1988). Bradley and Heinz-Joachim (1985) summarized that NLP is limited by a lack of supportive empirical evidence and is too simplistic to account for verbal behavior adequately.

Finally, in a meta-analysis and a follow-up investigation of controlled studies examining the effectiveness of NLP in influencing others, Sharpley (1984, 1987) stated that matching representational systems to eye movements has no effect. Similarly, Druckman and Swets (1988), in their evaluation of NLP for the National Research Council, concluded “there is little or no evidence to date to support either NLP assumptions or NLP effectiveness” (p. 143). The most telling commentary on NLP may be that in the latest revision of his text on enhancing human performance, Druckman (Druckman and Bjork, 1991) omitted all references to Neurolinguistic Programming.

Concluding Remarks

A key goal of HRD is to optimize individual, team, and organizational performance (Gilley and Eggland, 1989). To survive and prosper, organizations are interested in enhancement of human performance and are willing to pay handsomely for a competitive advantage. This is reflected in the American Society for Training and Development’s finding that employers in 1995 (most recently available figures) spent $55.3 billion on HRD activities in the United States, an increase of 20 percent (adjusted for inflation) over 1983 (“Trends,” 1996). With such funding available it is inevitable that a multitude of approaches will be offered to businesses. Some of these techniques have been and will continue to be developed outside of mainstream, scientific research contexts. And frequently these methods are accompanied by claims of extreme effectiveness.

We identified four alternative training techniques that have been widely touted and sold to government and industry: subliminal stimulation, mental practice, meditation, and NLP Finding that the claims made for three of these techniques, mental practice being the exception, did not stand up to scientific scrutiny, we encourage HRD professionals to guard against substituting testimonials and popularity in the marketplace for research evidence when they consider a new training aid. Effective, professional HRD practice depends on having and using objective research evidence to guide evaluation. One should question how practice, and the field of HRD, can advance when such training techniques are used if the research is not present.

As is the case with any product that comes with extraordinary claims, it is the buyer’s responsibility to be skeptical and cautious when weighing the effectiveness of nontraditional training and development products. As we indicated, some products or techniques may be useful, if in limited contexts, while others may be of little value. The only way to discern which is which is by using the scientific method and relying on formal experimental methodology instead of the testimonials or marketing hype typically used by proponents of alternative training and development strategies.










Queen’s University, Kingston, Ontario, Canada


Communication analysis has championed the existence of innate unobservable processes which influence communication, but it is limited by the lack of empirical research and the failure to recognize the importance of environmental influences on behavior. As an alternative to the narrow approach of behaviorism it is welcome, but in many ways the innate mechanisms of their neurolinguistic programming cannot account for all verbal behavior. Whether or not they are interested in efficacy or, simply, efficiency, experimental examinations will help to clarify their model and its effectiveness, and may lead to future improvements.

The theoretical approach to neurolinguistic programming designed by Bandler & Grinder (1975) was an attempt to uncover the structures inherent in all human behavior. The introduction of transformational grammar to the therapeutic process may be regarded as their main contribution to the field. Borrowing extensively from Chomsky’s (1957) theory of transformational grammar, Bandler & Grinder outline a step-by-step analysis of a client’s language as presented in therapy, the “surface” and “deep” meanings of these structures, and specific clinical techniques for dealing with them.

In an attempt to improve upon the paucity of background information supplied in their basic readings (1975, 1976), as well as to provide a critique of communication analysis, the historical context out of which this theory emerged will be examined. The knowledge gained by the historical analysis will provide some insight into the theory’s underlying premises.

Historical Background

Many of Bandler & Grinder’s theoretical assumptions can be traced back to Husserl’s phenomenological philosophy. In ways that are similar in structure to Husserl’s philosophy, Bandler & Grinder discuss the issue of distinguishing between the “real world” and the “perceived world.” Both argue that one’s perceived world represents only a limited model of the real world; our perception is influenced not only by neurological restrictions but also by social and individual histories.

In turning inward to examine the self, Husserl argued that the same neurological, social, and individual constraints affect our perception. The “I” which each of us is aware of is a subjective self (similar to Bandler & Grinder’s surface structure). Beyond this “I” lay a deeper hidden “I myself” (Bandler & Grinder’s deep structure). These concepts provided the philosophical basis for later linguistic investigations (White, 1955).

At the same time as Husserl’s philosophy was emerging, Wilhelm Wundt was working, in the field of psychology, with notions similar to those of Husserl. In his Voelkerpsychologie, Wundt examined the study of language, myth, and custom. He believed (as Bandler & Grinder do) that the experimental psychology of his time explored only the “outworks” of the mind, his Voelkerpsychologie was an attempt to reach deeper, hidden levels. The similarities between Wundt’s theory and Bandler & Grinder’s (1975,1976) description of deep and surface structure are easily seen. In addition to the theory of verbal expression, Wundt (like Bandler & Grinder) recognized the importance of nonverbal communication–gestures, expressive sounds, and primitive forms of expression.

Modem psycholinguistic ideas are clearly mirrored in Wundt’s approach; his concepts, however, were not widely incorporated by other researchers of that era. Instead, the American schools of functionalism and its offspring, behaviorism, dominated the psychological field. Linguistics, and its significance to psychology, was temporarily submerged.

Behaviorism remained at the forefront of experimental psychology until the 1950s, when theorists began to make serious attacks against its restrictiveness. Many limitations of behavior theory were discussed, but the critical insufficiency identified was its inability (in principle) to allow for symbolic processess, including meaning.

Along with this rise in “mentalistic” experimental approaches, two other attacks on behaviorism were occurring which subsequently had a strong influence on communication analysis. These critiques were spearheaded by Carl Rogers and Noam Chomsky.

Rogers (1951) was the major spokesman for a phenomenological psychology which emphasized empathy and the study of subjective experience.

A more strident and powerful criticism was directed at behaviorism by linguist Naom Chomsky. In 1959 he pointed out that behavioral terms such as stimulus, response, and reinforcement can be well defined in animal learning studies, but they undergo serious modification when extended to human behavior. Unlike Rogers, who called for a modification of behaviorism, Chomsky believed that it had to be replaced.

In its place, Chomsky (1957) created transformational grammar, a system which describes the grammatical structure of language. Although each new sentence is as creative as it is different, this element of creativity can be understood only through the recognition that language is a “rule- governed behavior.” People depend upon an innately determined set of grammatical rules in order to generate and understand verbalizations. Chomsky believed that verbal behavior could not be understood through the analysis of speech and hearing (as the behaviorists had attempted to do); a close analysis of these inner rules and underlying mental structures was necessary. Accordingly, human communication was divided into two levels, which can be roughly defined as 1) the “deep structure,” representing the message or meaning communicated by a sentence, and 2) “surface structure,” the syntactical form of the communication.

Bandler & Grinder’s theory has strong connections with both Rogers’ phenomenology and Chomsky’s transformational grammar. Together with Rogers, they focus on the client’s subjective experience, the need to understand this subjective world in order to induce change, and the desire to develop a systematic approach for achieving this. But, unlike Rogers, who was willing to accept that one mode of knowing was the objective mode and that behaviorism could be built upon, Bandler & Grinder seem to reject not only behaviorism but also its method of empirical research. Their interest is not in scientific evidence but in “what works.” With Chomsky, Bandler & Grinder share the belief that communication is rule-governed behavior which conforms to particular linguistic principles. Basing their theory on the work of Chomsky (1957), Bandler & Grinder have hypothesized about the internal mechanisms which generate language and psychopathology, as well as the means by which successful therapy is achieved.

Critical Evaluation

Given that communications analysis emerged in response to the challenges made to behaviorism by Rogers and Chomsky, and given that it is based, in part, on these theorists’ ideas, the work of Bandler & Grinder can be evaluated on the basis of its ability to meet those longstanding challenges. Since the need to understand human communication is central to both Bandler & Grinder’s theory of neurolinguistic programming and the rejection of behaviorism, this will be central to the evaluation of the theory.

The study of human communication is commonly divided into the three areas: syntactics, semantics, and pragmatics (cf. Watzlawick etal., 1967). Syntactics examines the problem of transmitting information–coding, channels, noise, redundancy, and the grammatical arrangements of words in speech. Semantics is concerned with the issue of the message, the sign’s meaning. The latter can be distinguished from syntax since it is quite possible to transmit strings of signs which are syntactically accurate, but semantically meaningless (Lewis Carroll’s “Jabberwocky” is perhaps the most famous example). Signs remain meaningless unless both sender and receiver agree beforehand on their significance. The effect of communication on behavior is the domain of pragmatics. Watzlawick et al. stresses that pragmatics is concerned not only with the effect of a particular communication on the receiver but also with “the effect the receiver’s reaction has upon the sender” (Watzlawick et al., 1967, p. 22). For this reason, the study of communication should focus on the sender-receiver relationship–their interactional pattern as mediated by communication–rather than on sender or receiver alone. The three areas, syntactics, semantics, and pragmatics, are clearly inseparable and interdependent. Superimposed on these three aspects of communication is the context in which the communication occurs. As stated by Watzlawick et al. (1967):

Failure to realize the intricacies of the relationship between an event and the matrix in which it takes place, between an organism and its environment, either confronts the observer with something ‘mysterious’ or induces him to attribute to his object of study certain properties the object may not possess (P- 21).

The behavior will remain unexplained as long as its context is not included. Any system of formal grammatical principles and innate structures fails to account for all aspects of human communication since the contribution of learning to language is ignored. As stated by Osgood (1968):

There is much about any particular language which is arbitrary and must be learned. Certainly this applies to meaning. The elaborate rules of reference (the class of events correlated with the usage of a particular term), of signification (the features which define membership in reference classes and hence project to new instances) and of implication (the nondefined but correlated features) must be learned; to put it crudely, there is nothing innately horsey about the word horse. The same thing applies to syntax. Although there are undoubtedly innately determined universal in language, particularly with regard to deep structure, languages differ markedly in the ways in which deep structure is related to surface structure, and these rules must be learned if the child is going to understand and create well-formed sentences in his particular language. The point here is this: Transformational grammars per se have nothing to say about learning; when they are treated as models for the language-user; they will be necessarily to that degree insufficient (p. 503).

Bandler & Grinder’s theory of neurolinguistic programming focuses on syntactics and semantics but virtually ignores pragmatics. If pragmatics are considered at all, it is only in a very limited way. Bandler & Grinder assess the client’s communication skills at the level of action and reaction but never move into the metacommunicative level of transaction. Watzlawick and his colleagues pointed out that a person does not originate communication, he or she participates in it–person A’s effect on В is only part of the story; B’s response will also affect A’s next move. From this perspective, communication analysis only scratches the surface of verbal behavior. To date, Bandler & Grinder as well as their supporters have made no attempts to incorporate the pragmatics of communication into their theory. This is a serious shortcoming which may limit the effectiveness of this approach.

A second aspect of communication which Bandler & Grinder have yet to consider is context. The ability or inability of an individual to understand and respond appropriately to the context of a verbalization is commonly associated with mental health (cf. Neale & Oltmanns, 1980). Verbal behavior that is out of context or “that shows certain other kinds of randomness or lack of constraint, immediately strikes us as much more inappropriate than merely syntactical or semantic errors in communication” (Watzlawick et al., 1967, p. 36). Also, picking up on contextual cues enables the individual to evaluate, influence, and predict verbal and nonverbal behavior.

Communication analysis, as presented by Bandler & Grinder, therefore, fails to account for both the pragmatics and the context of communication. As a theory, it brings us part of the way toward understanding this behavior, but the most intricate and important aspects have not yet been explored.

Could communication analysis be expanded to include context and pragmatics? It would seem impossible as long as Bandler & Grinder refuse to acknowledge the need for empirical research. Issues such as the way in which pragmatics and context are related to psychopathology are too complex to be studied by clinical observation alone. As previously stated, their theory is not clearly articulated; its terminology, premises, and assumptions are either specified in an ambiguous manner or not specified at all. Before any expansion can take place, the present state of the theory must be improved.

Bandler & Grinder sidestep the issue of research by stating that they are not interested in discovering the “truth,” only “what works.” It has yet to be empirically demonstrated that their approach works. As an example, Bandler & Grinder offer no scientific support for the connection which they assume exists between a client’s distorted perceptions of the world and his/her distorted linguistic representations. This supposition is central to their theory, yet they have not empirically tested it. Intuitively, one might agree with their assumption but in having to refer constantly to the intuitions a native speaker has in assessing the “well-formedness” of a clients’ messages, and in having to accept on faith the relationship between distorted perceptions and expression, it is difficult to accept that one is doing controlled therapy.

Bandler & Grinder have provided only personal testimony of their theory’s accuracy–and as they themselves state, personal perceptions are not always an accurate representation of reality.

Also in regard to this criticism, Bandler & Grinder discuss such linguistic concepts as deep structure and surface structure as if they existed within the individual. Even Chomsky (1957), however, was willing to recognize that an adequate model of language is not necessarily an adequate model of the person. Once again, it is up to Bandler & Grinder to demonstrate that their model of language is functional in terms of human communication.

In spite of these weaknesses, if we consider their approach as a practical guide to common linguistic distortions rather than a comprehensive, theoretical, linguistic analysis of clients’ messages, we might be able to gain some useful intuitions from their material.








Observations Concerning Research Literature on Neuro-Linguistic Programming

Eric L. Einspruch and Bruce D. Forman Department of Psychiatry and Department of Educational and Psychological Studies

University of Miami

There is a growing body of empirical literature on Neuro-Linguistic Programming (NLP). A review of this literature by Sharpley (1984) failed to consider a number of methodological errors. In the present article the authors identify six categories of design and methodological errors contained in the 39 empirical studies of NLP documented through April 1984. These categories include (a) lack of understanding of the concepts of pattern recognition and inadequate control of context, (b) unfamiliarity with NLP as an approach to therapy, (c) lack of familiarity with the NLP “meta-model” of linguistic communication, (d) failure to consider the role of stimulus-response associations, (e) inadequate interviewer training and definitions of rapport, and (f) logical mistakes. Representative reports reflecting each category are discussed. Suggestions are offered for improving the quality of research on NLP.


Neuro-Linguistic Programming (NLP) is a way of organizing and understanding the structure of subjective experience and is concerned with the ways in which people process information but not necessarily with the specific content of that information. Information is processed primarily in three modes: visual, auditory, and kinesthetic. The sensory modalities used in a given task and their sequence are critical to the performance of that task. Persons who are extremely skilled at a task will have radically different processing sequences than those who perform poorly on that same task. Understanding the structure by which the skilled person processes information, through the observation of eye scanning patterns and linguistic patterns, allows programs (similar to computer programs) to be codified, which can be taught to other persons (Dilts, Grinder, Bandler, Cameron-Bandler, & DeLozier, 1980; Kins- bourne, 1974).

A note of gratitude is extended to Amber Б. Goldstein, whose knowledge and continual support made this article possible. Acknowledgment is also given to Robert H. Dolliver for his editorial assistance in the preparation of this article.

Requests for reprints should be sent to Bruce D. Forman, Counseling Psychology Program, P.O. Box 248065, University of Miami, Coral Gables, Florida 33124.

Developed in 1975 by Richard Bandler, a mathematician, and John Grinder, a linguist, NLP has been clinically demonstrated as a powerful technology for engendering change (Bandler & Grinder, 1979; Grinder & Bandler, 1981). From their studies Bandler and Grinder developed skills of modeling that allow one person to identify in a specific fashion the structural elements of another’s behavior and to teach that structure to yet a third person (Dilts et al., 1980). When modeling another person the modeler suspends his or her own beliefs and adopts the structure of the physiology, language, strategies, and beliefs of the person being modeled. After the modeler is capable of behaviorally reproducing the patterns (of behavior, communication, and behavioral outcomes) of the one being modeled, a process occurs in which the modeler modifies and readopts his or her own belief system while also integrating the beliefs of the one who was modeled. Because they are extraordinarily skilled at this way of learning, Bandler and Grinder were able to ferret out the essential patterns used by Milton Erickson, Virginia Satir, Fritz Peris, and others, to codify these patterns in a succinct, understandable way, and to transfer the skills to others who are interested in learning them.

Gregory Bateson (1972), whose work forms a foundation for NLP theory, postulated four logical levels of learning. The first level is the level of content, and this is the level at which most people spend their lives. Here one learns how to tie one’s shoes, cook a meal, drive a car, and so on. Some people become acquainted with second-level learning: the learning of context, or learning how to learn. People who operate at the second logical level of learning may rapidly learn any new content-specific area, because they are capable of moving through the learning process in an efficient, effective manner. In rare cases, persons may rise to the third logical level of learning, the learning of how to learn context. In this case one is operating at a level of contextual pattern recognition; one is able to easily identify and operate on the structure of any experience. It is at this level that Bandler and Grinder operate when they are modeling (or teaching modeling to) some one. Bateson reserved his fourth class of learning for those accomplished persons like yogis and Zen masters.

With the foregoing as a conceptual framework for understanding NLP, we may turn to the consideration of the current research literature on NLP. The only current review of this literature was made by Sharpley (1984). Sharpley’s review is a reasonably thorough summary of some published articles on NLP, and he is to be commended for his efforts. The authors of studies he reviews make fundamental errors by neglecting the NLP model of pattern recognition, linguistic communication, and therapeutic intervention. In addition, these authors focus on the primary representational system (PRS) and reify the term, another major mistake. The danger of reifying terms is that one may be easily led to mistake a construct for reality (Kor- zybski, 1941). Sharpley does not address these issues. Other limitations of his review are the statement that PRS patterns occur for right-handed people only (not mentioning the importance of calibration to the individual, i.e., recognizing individual differences in patterning) and the jump from the deep structure/surface structure of language to representational systems without indicating how and why this concept is a crucially important one (e.g., citing seminal works by Noam Chomsky). Nowhere do Bandler and Grinder suggest that matching the client’s PRS is the key to effective counseling, as Sharpley states, but rather they indicate that it is an important element in effective communication and should be used in conjunction with other techniques. Sharpley reports that the amount of published data supporting NLP as a viable model for therapeutic change is minimal. Nevertheless, many skilled NLP practitioners have a wealth of clinical data indicating that this model is highly effective. Clearly these practitioners would provide a service to the field by presenting their data in the literature so they may be critically evaluated.

In preparing this article the authors identified 39 reports of empirical studies by searching journal citations and several automated data bases of publications from 1975 to April 1984. Twenty-four of these articles were not reviewed by Sharpley. Almost all of the research articles concern either the identification of primary representational systems or the effects of matching or mismatching primary representational systems. The following discussion considers these categories of design errors: (a) lack of understanding of concepts of pattern recognition and inadequate accounting of context, (b) unfamiliarity with NLP as an approach to therapy, (c) unfamiliarity with the NLP “meta-model” of linguistic communication, (d) failure to consider the role of stimulus-response associations, (e) inadequate interviewer training and definitions of rapport, and (f) logical mistakes. Representative studies that include these errors will be discussed. The studies selected appeared in the journal literature and clearly represent the problems discussed.

Design and Methodological Errors

Lack of Understanding of the Concepts of Pattern Recognition and Inadequate Control of Context

Neuro-Linguistic Programming is based on the identification and interruption of limiting behavioral and cognitive patterns and the generation of more useful and appropriate ones. Gumm, Walker, and Day’s (1982) study demonstrates limited understanding of these concepts (in addition, they misspell the name of the topic, consistently referring to “neurolinguistics” programming). In this project 50 undergraduate women were interviewed for 45 min each. Responses to a set of questions were categorized by frequency of each representational system. Subjects were also asked to complete a self-report questionnaire of what they thought were their primary representational systems. After the interview, subjects were led into a room surrounded by curtains and were placed in a chair where they had their heads secured in a restraining device. They were then asked questions and were rated on eye scanning patterns. An analysis was made between each of the three methods of determining primary representational systems, and no agreement was found, although each method appeared to be biased toward one of the sensory modalities.

Gumm et al. (1982) selected only right- handed people, stating that Bandler and Grinder claim that their model does not work for left-handed people. Bandler and Grinder make no such claim; rather they state that most people follow a standard pattern of eye accessing cues. In every case, they emphasize the importance of calibrating to the individual being interviewed, understanding that there is no substitute for paying sufficient attention to gathering sensory-based data. Calibration is the process by which one tunes himself or herself to the nonverbal signals that indicate a particular state in a particular person, presupposing that the meaning of a communication is the response it elicits (Grinder & Bandler, 1981). Right-handed people will occasionally violate the usual pattern, and many people have partial or full reversals of this pattern. The salient feature of the concept of patterning is that whatever pattern a person may have, he or she will follow that pattern consistently. It is a major failure for the authors not to perform individual calibrations.

Patterns of cognition may occur at either a conscious or an unconscious level In neither this study nor any of the others do the authors make a distinction between what is in awareness and what is outside of it (Erickson’s “conscious” and “unconscious” minds). It is quite common for a person to spend most of his or her time accessing (by eye movements) in one representational system and speaking (representing in linguistic form) in another representational system. This is simply a statement about what is and what is not in the speaker’s awareness. All information not considered in a verbal communication, unless otherwise investigated, may be considered to be outside of the person’s awareness. Understanding this concept and what to do with the information obtained is an important skill in NLP practice. Finally, it is necessary to understand that the representational system in which information will be stored or from which it will be retrieved is highly contextualized (i.e., varies with the situation), and this context will directly influence the system used. Locking a person’s head in a restraint while asking questions that are non sequiturs cannot not affect how the subject will respond. Context plays an important role in determining the meaning as well as the structure of any communication. Thus, the results of this project must be called into question. These mistakes are also found in Beale (1980/1981), Birholtz (1981), Cole-Hitch- cock (1980), Fromme and Daniell (1984), Hernandez (1981), Johannsen (1982), Kraft (1982), Lange (1980/1981), Matter (1980/1981), Owens (1977/1978), Radosta (1982), Shaw (1977/1978), Talone (1983), and Thomason, Arbuckle, and Cady (1980).

Failure to Understand NLP as an Approach to Therapy

Neuro-Linguistic Programming theory presupposes a particular approach to therapy that was not followed by the researchers in the works that have been reviewed. For example, Hammer’s (1983) interview was supposed to represent a counseling situation with interviewer questions designed to elicit representations of past experiences. This is a distortion of the NLP approach to therapy, which is generative by nature and makes use of questions designed to build a future that is appropriate for the client to move toward.

Yapko (1981a, 1981b) has probably produced the most sensible NLP research project in the current literature, if only because he used an operational, physiological indicator of rapport. However, he does not control for contextual influences. Thirty graduate students in counseling were inducted into hypnotic trance three times. Each time the trance was induced with a different representational system emphasized. The inductions were taped for standardization. Depth of trance was measured physiologically by electromyographic recordings of muscle relaxation. Between inductions participants were asked to rate their subjective feelings of relaxation. The main variable not controlled was the effect of moving into and out of trance states. Milton Erickson frequently used alternate entry into and out of trance as a means for deepening the trance state (Grinder, Bandler, & DeLozier, 1977). Yapko’s results support Bandler and Grinder’s (1979) contention that matching primary representational systems enhances rapport. Subjects achieved greater states of relaxation, both subjectively and as measured physiologically, when the induction was. presented in a matched system versus a mismatched system. Unfortunately, the uncontrolled influence of context limits the interpretation of the results.

Allen (1982) performed the only study that used behavioral observation, although he incorporated other design errors. Neuro-Linguistic Programming is noted for its rapid treatment of simple phobias (as distinct from social phobias or agoraphobia), and Allen studied NLP versus Massed Systematic Desensitization (MSD) versus no treatment with 36 undergraduates with snake phobias. No differences between the NLP and MSD groups were found on posttreatment behavioral approach tests. In other words, the NLP- based treatment was just as effective as the MSD treatment. Despite the equivalence of findings, those in the NLP group reported more frequently that they thought that they were over their fear of snakes. The study at least follows the NLP model in that it uses observable outcomes as dependent variables. Other studies that failed to understand the NLP approach to therapy include Atwater (1983), Ellickson (1983), Hagstrom (1981/1982), Haynie (1982/1983), and Wilimek (1979/1980).

Failure to Understand the Meta-Model of Linguistic Communication

Ellickson (1983) selected 72 college students and determined each student’s primary representational system by observation of eye scanning patterns. Subjects were then randomly assigned to one of two interview conditions in which the interviewer either deliberately matched or mismatched predicates presented by the subject. After a 15- to 20-min interview, subjects rated their interviewer on perceived qualities of empathy, ease, anxiety, and hostility on three self-report instruments. Hypotheses related to the effects interactions of sex of interviewer, sex of subject, and interview condition were tested. Only sex interactions produced significant differences, and little evidence was generated to support Bandler and Grinder’s (1979) claim that matching representational systems enhances rapport.

Ellickson’s failure to consider the NLP meta-model of linguistic communication (Bandler & Grinder, 1975) is a major flaw in this study. The meta-model is a tool used to understand how a person transforms the reference structure of experience into the surface structure of language by classifying the syntactical components of language. This is a primary NLP tool because it is essential for eliciting the structure of how a person constructs his or her reality and its limits and it provides a way to engage a person in conversation that accesses information and resources and directs them toward a therapeutic outcome. In the NLP model, words like empathy, ease, anxiety, and hostility are all nominalizations (i.e., nouns that are generated by the reification of verbs). As such, they do not represent any tangible thing but are instead the products of the constructs of each individual, predicated on the individual’s strategy for constructing reality. Helping the client to turn nominalizations back into action words is an important NLP technique performed by appropriate use of the metamodel. Consequently, there is no way that any of these words can be measured, either for a group or particularly by a paper-and- pencil self-report inventory. These mistakes are also found in Frieden (1981) and Pantin (1982).

Failure to Consider Stimulus-Response Associations

Dorn (1983a), who was concerned with accurately assessing a client’s primary representational system, designed a study attempting to determine the best method for doing so. Unfortunately, he did not consider the nature of stimulus-response associations. Three methods were used. First, subjects were introduced to a relaxation exercise and then were asked to verbally describe three different scenes. Already context is specified and not generalizable. Next, subjects were presented with a word list containing 18 triads, each triad containing one visual, one auditory, and one kinesthetic word. The subjects were instructed to pick the preferred word in each triad. The one example provided in the text contained a nonspecified predicate, so one must question the choice of words in the other triads. Also, Dorn disregarded the NLP concept that words are simply anchors and the subject was likely to respond to the most positive anchor rather than respond to a word because it represented a particular sensory modality. In this case, the positive anchors were words with which the individuals had positive associations. Finally, subjects were given an explanation of the concept of primary representational systems and were asked to self-report which one they thought was theirs. This presupposed a conscious awareness of how one represents and maps one’s world, a process that is necessarily unconscious (Bateson, 1972). Predictably, analysis of the data revealed no significant results.

Inadequate Interviewer Training and the Nature of Rapport in the Counseling Relationship

All of the reviewed studies failed to provide adequate investigator training. For example, in Dowd and Hingst’s (1983) study, master’s level students who had no experience as therapists were trained in four 90-min sessions. This may allow enough time for a person to learn to identify eye accessing cues, but it does not provide enough time to develop mastery of the NLP framework for establishing rapport. Rapport is not just a function of matching or mismatching representational systems. It is a complex matter involving matching at many levels, including the pacing of the client’s breathing and many elements of linguistic structure. Rapport may be defined operationally as occurring when the client is willing to follow the lead of the therapist. An interviewer may match representational systems and still mismatch in enough other ways that rapport is not established. Similar errors are also found in Appel (1983), Brockman (1980/1981), Cody (1983), Ellickson (1983), Dorn (1983b), Dowd and Pety (1982), Ehrmantraut (1983), Falzett (1981), Green (1979/1981), Hammer (1983), and Paxton (1980/1981).

Logical Mistakes

Mercier and Johnson (1984) conducted a study based on an interesting idea: taking classic counselor training films and analyzing them from part of an NLP perspective. However, the authors made some inappropriate logical jumps: They predicted how the therapists in the film would use predication patterns in their own language and predicted whether the therapists would match or mismatch client predication on the basis of the system that the therapist followed. The system that the therapist subscribes to will be in part determined by his or her way of understanding the world, and predication patterns partially reflect the structure of that understanding. Mercier and Johnson’s predictions were in the wrong direction; the model may or may not be predictable from the predication patterns, but certainly the reverse cannot hold true. Once again, the investigators have failed to calibrate to the individuals being studied. These mistakes are also included in Ellis (1980/1981) and Frye (1980).

Conclusion and Recommendations

Neuro-Linguistic Programming is an extraordinarily complex model of human cognition and behavior and of how to identify behavioral and communication patterns and interrupt these patterns in a deliberate way so as to achieve predictable outcomes. Eye scanning patterns and representational systems are an important, but small, part of NLP. It is difficult to understand the NLP framework from the perspective of traditional counseling models; it is much more appropriate to approach it from the framework of mathematics, biology, or cybernetics. Neuro-Linguistic Programming deals with patterns of interactions, and to ignore this basic premise is to miss an essential feature of NLP as a model of understanding and altering human behavior.

A number of modifications could be made to improve designs of research conducted on NLP. First, and perhaps most important, researchers should be trained by competent NLP practitioners for an appropriate duration of time. Training should include pattern recognition skills and a foundation in the presuppositions of NLP to provide an adequate framework for understanding NLP as an approach to therapy. Second, any investigators wishing to study rapport should rely on objective, sensory-observable measures of their procedures. Third, except in gathering information about sensory-observable outcomes of therapy, procedures should not be applied at the group level, but rather at the individual level, and calibrated to each person’s personal characteristics. Data on individual observations may then be summed to arrive at group data. Fourth, any treatment studies should be performed only by therapists with demonstrated mastery of the model and its techniques, and outcome measures should be behavioral in nature. If comparisons are to be made with other treatment approaches, the therapists using the comparative model should be equally proficient. For example, the treatment of phobias with NLP is a specialized operation, because not every phobia is to be treated with the same techniques or for the same duration. Neuro-Linguistic Programming is a complex model requiring extensive training before a practitioner may legitimately undertake a study of this nature. One cannot simply attend one or two workshops or read a book and assume that he or she can effectively perform NLP therapy any more than this could be assumed for any other model of therapy. Given the current state of research on NLP, a more appropriate strategy would be to pose questions concerning the sensory-observable outcomes of therapy rather than investigate pieces of the model and their relations to reified concepts.

In conclusion, on the basis of the research that has appeared in the literature, it is not possible at this time to determine the validity of either NLP concepts or whether NLP-based therapeutic procedures are effective for achieving therapeutic outcomes. Procedures generated from the NLP model must be used within the presuppositions of the model, and research on reified concepts is trivial in nature and is a distraction from the serious issues relating to testing the NLP model. Only when well-designed empirical investigations are carried out may we be assured of NLP’s validity as a model of therapy.








This chapter is a modified version of a paper given by the author at the Fourth European Congress of Hypnosis in Psychotherapy and Psychosomatic Medicine, 1987, Oxford.

Neurolinguistic programming (NLP) is a model of human behaviour and cognition which describes how people represent their world, how they interact and communicate with it and with one another, how it can be that they can experience distress and disappointment in these interactions, and how they can be helped to change their representation of the world to alleviate their distress and cope with life more effectively and with greater fulfilment. Based on the tenets of NLP, strategies have been formulated whereby it is asserted that counsellors,therapists and communicators may enhance their effectiveness in helping their clients, and therapeutic procedures have been outlined which it is claimed bring about far more rapid and effective changes than hitherto in the formal practice of psychotherapy.

Perhaps the central philosophy of NLP is most aptly summed up in the sentence ‘The Map is not the Territory’ (see, e.g. Lankton, 1980, p.7). That is, each one of us only ever operates on the basis of our internal representation of the world (our ‘map’) and not the world itself (the ‘territory’). The maps thal we create may be limited in many ways, impoverished, distorted and inflexible. The choices which we thus make available to ourselves are restricted, and our transactions with the world will accordingly,be frustrating and difficult. It is therefore the therapist’s task to understand and operate on the basis of the client’s map of the world in order to assist the client to overcome these restrictions and thus provide him with more choices.


One of the important concepts of NLP is the primary

representational system (PRS). The maps that people make of their world are represented by the five senses, visual (V), auditory (A), kinaesthetic (K), olfactory (0) and gustatory (G). V, A and К are thought to be the major ones and individuals differ in the way they employ these representational systems.

For example, a person may tend to represent his world in the V mode, i.e. through internal pictures, another person may tend to use a К representation, i.e. through feeling, and a third person may have a predominantly A representation, through sounds and verbalization (Grinder and Bandler, 1976; Bandler and Grinder, 1979).

The personal consequences of having one PRS as opposed to another are not greatly elaborated upon. (Relationships are reported involving preferred leisure activity (Frye, 1980) and degree course (Ellis ,1981) . What ±s_ stressed, however, is that it is advantageous for the therapist to ascertain the client’s PRS.

How is this achieved? Firstly it is claimed that at any time the representational system being employed is revealed by aperson’s style of speaking, specifically in the predicates he uses — verbs, adjectives and adverbs. A person with a V PRS will tend to use predicates such as ‘I see…’, ‘It appears to me’ and ‘I have a clear picture…; someone with an A PRS will use phrases such as ‘I hear…’, ‘It sounds to me..’ and ‘I tell myself…’; and expressions such as ‘I feel’, ‘He’s out of touch’ and ‘It’s heavy going’ will be favoured by someone with a К PRS. A second indication of representational system is direction of eye movement. It is claimed that a person accessing V information will tend to look upwards (left for remembering, right for constructing); a person looking horizontally left or right will be accessing A information (remembered and constructed, respectively), likewise looking downwards and to the left; a person looking down and to the right is accessing К information, and a final eye position is eyes unfocused and looking ahead which is interpreted as accessing visually represented information (Lankton 1980, p. 46).

Matching of Primary Representational Systems NLP writers contend that by matching, mirroring or pacing the client’s verbal and non-verbal behaviour,

i.e. matching aspects of speech, gestures, body posture, breathing, blinking, etc, one is tuning in on the client’s representation of the world and thereby facilitating rapport, understanding, trust, communication and so on. NLP writers are also emphatic that by matching the client’s PRS — i.e. using predicates in the same mode — rapport and therapist effectiveness will be considerably enhanced. Conversely, mismatching the client’s PRS will impede communication, lead to misunderstandings, loss of rapport, and resistance (see, e.g. Bandler and Grinder, 1979, p.ll). Consequently, according to the NLP model, to enhance one’s effectiveness as a communicator one must establish the other person’s PRS and match one’s predicates, as well as other verbal and non-verbal behaviours, with those of the clients.


Most experimental research into NLP has investigated the claims regarding the concept of PRS, notably the following three assertions:

  1. The hypothesis that a person has a PRS which is observed in his choice of predicates: A number of studies have investigated this hypothesis by attempting to identify groups of people distinguished by their preferences for V, A or К predicates in their speech on simple interviewing or direct questioning (see Grinder and Bandler, 1976, p.12, for questioning procedure).

Beale (1981), Lange (1981), Dorn (1983a), Petroski (1985) and Coe and Scharcoff (1985) obtained no clear distinctions between groups of subjects in terms of a PRS based on predicates. Birholtz (1981), Falzett (1981), Johannsen (1982), Gumm, Walker and Day (1982), Graunke (1984), Faulkender (1985) and Ridings (1986) all report that a large majority of Ss have a preference for using К predicates. Birholtz (1981) found that this preference (in all 27 Ss) was stable over time, although Ridings (1986) found the К preference to be less stable; in this investigation of 65 Ss, 55 showed a clear PRS on initial testing. This was in the К mode for 53 of these Ss, but the number of these showing a К preference dropped to 40 after 6 weeks. Finally, in Shaw’s (1978) study, none of a sample of 108 students could be identified as having a PRS in the V mode using verbal report, and none of Matter’s (1981) Ss had a predicate preference in the A mode.

Several investigators have explored the relationship between PRS, imagery material and imagery indices. Graunke (1984; see also Graunke and Roberts, 1985) gave Ss 10 imagery tasks associated with a particular modality (V, A, К or mixed). Ss’ predicates were directly related to the type of imagery; that is, Ss tended to switch from one mode to another according to the type of the imagery task. This does not support a trait concept of PRS. Beale (1981) also reported that predicate mode and material were related. Other negative findings regarding imagery and PRS were reported by Lange (1981), Johannsen (1982) and Fromme and Daniell (1984). For a discussion of strength of imagery preference and PRS see Grinder and Bandler (1976, p. 8-9).

Faulkender (1985) obtained no relationship between PRS in verbal report and performance on tasks involving the different perceptual modes (V, A and K). This author, however, reported some limited support for a classification of representational system using second and third preferences, and Pantin (1982) found a positive correspondence between S’s dominant mode of predicate usage and self reports of imaginal style.

In view of the preponderance of negative results concerning the hypothesized PRS as exhibited by S’s choice of predicates, it is worth noting some incidental positive findings that have been observed (but not yet replicated) by a number of investigators.

For example, it was noted above that Johannsen (1982) observed some positive correlation between predicates and certain indices of imagery. Birholtz (1981) observed that the proportion of A predicates used by Ss was correlated with measures of well-being, socialization, achievement via conformance, and intellectual efficiency on the California Personality Inventory. O’Leary (1984) has reported a significant correlation between S’s most accurate representational system as measured on the Affect Sensitivity Scale, and the thinking/feeling personality on the Myers-Briggs Type Indicator. Finally, Wilimek (1980) obtained some significant differences in predicate usage by high- and low-adjustment couples describing satisfying and upsetting experiences.

  1. The hypothesis that a person has a PRS which is observed in the direction of his eye movements : Investigators of this hypothesis have generally looked for correlations between perceptual processing (V, A or K) and ocular movement, and for consistent differences between individuals in preferred direction of gaze. It may be mentioned in passing that the early writers on NLP surprisingly make hardly any reference to the existing experimental literature on eye movement and cognitive mode. Even in the ’70s this work was quite extensive (see review by Ehrlichman and Weinberger, 1978). One of the early prominent researchers was Kinsbourne; in one study (Kinsbourne, 1972) he observed that when solving verbal problems (interpreting proverbs), right-handed Ss tended to turn their head and eyes to the right, whereas with problems of calculation and visualizing familiar places they tended to look upwards and to the left. These relationships did not hold for lefthanders. Although some of the findings on ocular gaze and cognitive mode may be compatible with the NLP model (e.g. Kinsbourne’s spatial task) by and large the proposals are not supported in such studies. No support for the NLP assertions on eye movements was found in investigations by Thomason, Arbuckle and Cady (1980), Beale (1981), Radosta (1982) , Cody (1983), Dorn, Atwater, Jereb and Russell (1983) , Petroski (1985), Farmer, Rooney and Cunningham (1985), Poffel and Cross (1985), Coe and Scharcoff (1985) and Elich, Thompson and Miller (1985). Cody (1983) moreover found little stability over one week for the very marginal eye-movement preferences exhibited by Ss.Beale (1981) noted a predominance of upward eye movements regardless of stimulus; such movements are, in NLP terms, associated with the V mode, but Poffel and Cross (1985) found in their Ss that the responses associated with V and A occurred equally often (22% for both, and 6% for K; 50% of the time no eye movements were recorded at all). Also Wertheim, Habib and Cumming (1986) found more responses associated with the A mode for all types of materials; nevertheless, in accordance with the NLP model, Ss showed a significant tendency to make upward eye movements for visual items and some support for this also comes from a study by Hernandez (1981).

No clear relationship has been reported between representational system as inferred by spoken predicates and by eye movements (Beale, 1981; Cole- Hitchcock, 1980; Gumm et al.,1982; Petroski, 1985; Coe and Scharcoff, 1985; Elich et al., 1985). However, Cole-Hitchcock (1980) did find a consistency in indicators for PRS in the V and A mode from written responses and eye movements. This was not observed by Coe and Scharcoff (1985). Finally, Owens (1978) reported that combining data from eye- movements with those from predicates in verbal reports gave a statistically reliable means of determining PRS. This was notreplicated by Gumm et al. (1982) who also checked Owens’s statistical result and found this had been incorrectly reported as significant.

  1. The hypothesis that communicators may enhance effectiveness if they match their client’s PRS in their choice of predicates; This hypothesis has been investigated by first observing the presumed PRS of Ss in their choice of predicates or in their eye movements and then subjecting Ss to an analogue counselling interview in which the counsellor is instructed either to use predicates which are congruent with Ss designated PRS or to deliberately mismatch the PRS. Occasionally a third condition in which there is neither matching nor mismatching (the ‘unmatched’ condition) has been employed. Ratings of interviewer qualities such as empathy, trustworthiness and attractiveness are then made by interviewees, interviewers and independent assessors using standardized scales. Sometimes the experimental Ss are the judges and they rate the quality of a given interview in which the different conditions of matching are manipulated.

In general, failure to confirm the hypothesis under discussion for such counsellor qualities as rapport, trustworthiness, perceived expertness and effectiveness, has been reported by Green (1981), Rebstock (1980), Frieden (1981), Dowd and Pety (1982),* Dowd and Hingst (1983), Dorn (1983b), Cody (1983), Schneider (1984) and Carbonell (1985). Moreover, Cody (1983) found that therapists who matched their client’s language were rated as less trustworthy and less effective. Also, an incidental finding of Frieden (1981) was that predicate matching did appear to produce more eye contact, but paradoxically increased head-to-head distance. Appel (1983) failed to find any effect of congruency in PRS on the attractiveness of speakers as perceived by male and female Ss, although a significant effect involving Ss’ secondary representational system when speaker and S were of the opposite sex was observed.

Support or partial support for the hypothesis comes from the following reports. Brockman (1981) found that interviewees rated a counsellor instructed to match predicates as more preferred and more empathic than one not so instructed; the empathy difference was also presented in judges’ ratings. However, there was no difference in willingness to self- disclose. Shobin (1980) found in an initial psycho- *Contrary to the references by Gumm et al. (1982) and Ellickson (1983), no positive findings for matching were reported by Dowd and Pety (op.cit.).

therapy interview that predicate matching gave higher ratings than a ’modified verbal pacing’ method in which other verbal elements (e.g. voice tone, tempo and syntax)were matched but predicates mismatched. Schmedlon (1981) reported superiority of matching over mismatching for empathy but no differences for perceived level of regard and depth/value, smoothness/ ease and positive feeling. Falzett (1981) reported a similar positive finding for perceived trustworthiness. In contrast to studies mentioned earlier, Pantin (1982) reported that Ss evaluated more positively recordings of simulated counselling in which client and counsellor were matched for dominant predicate mode. Ellickson (1981, 1983) obtained negative findings for the effect of matching on a number of scales, but ease of communication was superior in the matching condition for males only. Day (1985) found a ‘matching strategy’ (unspecified) gave superior scores for counsellors rated on expertness, attractiveness, trustworthiness and having utility. Paxton (1981) found no difference between matching and mismatching predicate systems on counsellor relationship but both methods were more effective than the no-matching condition. It was concluded that it was beneficial for counsellors to consistently use one representational system. Finally, Hammer (1983) has reported a favourable influence on perceived empathy of interviewer, if the latter tracks the predicates (i.e. matches them individually) of the interviewee rather than matches a single presumed PRS. Hammer concludes that PRS is not a useful concept but the predicates themselves have a perceptual significance because interviewers matched according to modality rather than responding with identical predicates. Other investigators who have used tracking rather than matching PRS include Ellickson (1981, 1983), Frieden (1981), Dowd and Pety (1982) and Dowd and Hingst (1983).

Several researchers have examined whether matching verbal material to S’s preferred predicates improves task performance or potentiates the effectiveness of the material. For example, Kraft (1982) explored the effectiveness of audiotaped relaxation instructions framed predominantly in the V, A or К modes. No benefits were observed when the instruction were congruent with S’s preferred predicate system. However, Yapko (1981a, 1981b) found that hypnotic relaxation instructions were increasingly effective when framed in the least to the most preferred mode, in line with NLP predictions. Talone (1983) found no evidence of a match in the mode of S’s responses (writing an essay) and mode of the experimenter’s suggestions (V, A or K). Shaw (1978) found no effect of matching material to be recalled to S’s preferred mode. Cody (1983) did not observe any expressed preferences due to matching amongst Ss for audiotaped vignettes concerning common pleasant experiences. However Pantin (1982) found that performance on a memory task was facilitated when mode of item presentation was congruent with S’s preferred mode of predicate usage, and Mattar (1981) found some evidence of greater ease of comprehension for segments of spoken material congruent with S’s inferred PRS (V or K).


In 1984, Sharpley reviewed the experimental evidence for the tenets of NLP relating to the PRS. He concluded that there was no support for the identification of the notional PRS by predicates and eye movements and no consistent evidence for the alleged benefits of predicate matching in counselling. Dorn, Brunson, Bradford and Atwater (1983) also concluded from their review of the literature that there was no demonstrably reliable method of assessing the hypothesized PRS. Sharpley’s review was criticized by Einspruch and Forman (1985) to which a rejoinder has been issued (Sharpley, 1987) in which the writer restates his position in the light of further evidence. The objections and counter-objections will not be fully discussed here and the reader is referred to the original papers, and to criticisms of research made by Beck and Beck (1984).

The present author is satisfied that the assertions of NLP writers concerning representational systems have been objectively and fairly investigated and found to be lacking. These assertions are stated in unequivocal terms by the originators of NLP and it is clear from their writings that phenomena such as representational systems, predicate preferences and eye-movement patterns are claimed to be potent psychological processes, easily and convincingly demonstrable on training courses by tutors and trainees following simple instructions, and, indeed, in interactions in everyday life. Therefore, in view of the absence of any objective evidence provided by the original proponents of the PRS hypothesis, and the failure of subsequent empirical investigations to adequately support it, it may well be appropriate now to conclude that there is not, and never has been, any substance to the conjecture that people represent their world internally in a preferred mode which may be inferred from their choice of predicates and from their eye movements.

These conclusions, and the failure of investigators to convincingly demonstrate the alleged benefits of predicate matching seriously question the role of such a procedure in counselling. It may be however that the general process of matching linguistic style and other verbal and non-verbal behaviours is of value, and this would still be consistent with NLP formulations. In accordance with this, some writers have suggested that with increasing familiarity there may be a tendency in counselling (and probably other) interactions for each participant to accommodate the other’s linguistic style — types of verb phrase, sentence length, and so on (Beiber, Patton and Fuhriman, 1977; Mercier and Johnson, 1984). Also, a test by Sandhu (1984) of the benefits of mirroring non-verbal behaviour (movements of extremities and posture directly, other movements indirectly) was positive for empathy although not for trustworthiness or positive interaction.


This verdict on NLP is, as the title indicates, an interim one. Einspruch and Forman (1985) are probably correct in insisting that the effectiveness of NLP therapy undertaken in authentic clinical contexts of trained practitioners has not yet been properly investigated. If it turns out to be the case that these therapeutic procedures are indeed as rapid and powerful as is claimed, no one will rejoice more than the present author. If however these claims fare no better than the ones already investigated then the final verdict on NLP will be a harsh one indeed.





Research Findings on Neurolinguistic Programming: Nonsupportive Data or an Untestable Theory?

Christopher F. Sharpley

Faculty of Education, Monash University, Clayton, Victoria, Australia

In an earlier review of the experimental literature on neurolinguistic programming (NLP), Sharpley (1984) drew the conclusion that the effectiveness of this therapy was yet to be demonstrated. In their comment on that review, Einspruch and Forman (1985) agreed with this conclusion hut suggested that it was due to the presence of methodological errors in the research on NLP to date and that the efficacy of NLP was open to debate. Unfortunately, those suggestions were based on misconceptions regarding the factors that limit the methodological worth of research. Several of the detailed criticisms from that review are refuted here, and further data from seven recent studies that further demonstrate that the research data do not support either the basic tenets of NLP or their application in counseling situations are presented. Implications from these findings for the use of NLP in counseling research or clinical practice are discussed.


Described by its founders as therapeutic magic (Handler & Grinder, 1975), neurolinguistic programming (NLP) suggests that the process of effective communication between persons (particularly counselors and clients) can be enhanced by identifying their “preferred representational system” (PRS) and by using this particular communication modality preferred by a client as a method of effectively helping that client to make changes in behavior. This process has been referred to as predicate matching and is a basic tenet of NLP (Bandler & Grinder, 1976, p. 8). In spite of a wealth of advertising that claims a great deal of clinical success in the use of NLP, there have been few reviews of the experimental literature in this area. In an earlier article (Sharpley, 1984), the outcomes of 15 studies identified in the literature were reviewed, and it was concluded, “At present, there is no consistent support for the use of predicate-matching in either contrived counseling situations or actual clinical realities” (p. 247).

In a reaction article to that review, Einspruch and Forman (1985) suggested that I had omitted criticizing the studies covered in my review on the basis of how well-informed the writers of those papers appeared to be on the various theoretical underpinnings of NLP. Einspruch and Forman then went on to gather 24 more studies and to criticize all 39 on six points relating to how well the various authors demonstrated their knowledge of certain intrinsic aspects of NLP in their articles. All 39 studies were then classified as not contributing to a reliable overall evaluation of the effectiveness of NLP, and the authors concluded, “It is not possible at this time to determine the validity of… NLP” (p. 594).

Einspruch and Forman (1985) should be commended for their efforts in finding these other 24 reports and for the issues

1 thank Allen E. Ivey. University of Massachusetts, for his helpful comments and suggestions on this article.

Correspondence concerning this article should be addressed to Christopher F. Sharpley, Faculty of Education, Monash University, Clayton, Victoria 3168, Australia.

that they raised in their review. To state, as they did, that it is not possible to draw any reliable conclusions about the effectiveness of NLP at this time deserves some comment If 44 studies (2 of those referenced by Einspruch & Forman were irrelevant because they did not directly assess a principle or procedure from NLP, and 7 more will be referred to below) of a particular procedure do not show any conclusive effects, then either there is great disagreement in the results that have been reported from these studies, or the procedure is not able to be adequately assessed. As I shall point out, research on NLP has consistently shown very few significant effects that lend support to claims of therapeutic magic (or even any degree of effectiveness at all), either in research designed to evaluate the basic principles of NLP or in the treatment of both “laboratory” and real clients.

Einspruch and Forman’s (1985) Criticisms of Research

Several points from Einspruch and Forman’s (1985) review require comment before going on to evaluate the extra literature to which they referred. First, they criticized the authors of the experimental literature for reifying the term PRS and for performing a series of studies to determine whether PRS exists and whether matching clients’ PRS assists in communication or leads to the more effective application of a variety of therapeutic procedures (e.g., relaxation training, recall of words, hypnotic induction, and treatment of phobias). There is little more that a researcher can do, however, to evaluate a theory than to test the veracity and strength of those principles of behavior that are held by the proponents of that theory. To evaluate NLP without testing for the presence of the PRS and its usefulness in communication and treatment appeals to be impossible, and investigation of those aspects of NLP is to be applauded rather than condemned as reification. Bandler and Grinder (1979) claimed, “People make movements with their eyes which will indicate to you which representational system they are using” (p. 18). Researchers may be excused for directing some efforts toward testing this claim.

Second. Einspruch and Forman (1985) criticized Gumm, Walker, and Day (1982) for claiming that NLP applies only to right-handed persons. Gumm et al. quoted this claim from a workshop run by Bandler and Grinder, and whether it is a widely accepted principle of NLP or not, it demands examination in controlled studies. It is not a criticism of the value of such studies that other proponents of NLP have not made this suggestion in favor of right-handed persons. It is, on the contrary, to be expected that researchers will try to set up experimental situations for testing the principles that underlie NLP. To date, there is little evidence to support those principles and much to suggest that they are invalid. (In fact, Bandler and Grinder, 1979, presented a diagram for identifying PRS from eye movements. They called this diagram “visual accessing cues for a normally organized right-handed person” [p. 25]. Additionally, they referred [pp. 21-36] to their system of PRS identification as being valid for right- handed persons and [perhaps] reversed for left-handed persons. Gumm et al. may, therefore, be seen as having followed Bandler and Grinder’s (1979) guidelines by restricting themselves to right-handed subjects and thereby maximizing the likelihood of finding an outcome that was supportive of this NLP principle.)

Another point for which Einspruch and Forman (1985) criticized the literature was the “failure to understand NLP as an approach to therapy” (p. 591). To demonstrate this point, Einspruch and Forman criticized Hammer (1983) for setting up an interview with questions designed to elicit the client’s PRS, calling this a “distortion of the NLP approach to ther- apy” (p. 591). Yet, in the NLP literature (e.g., Bandler & Grinder, 1976), the presence of the PRS is accepted as a fact, and its accurate identification is a primaiy aspect of effective therapy in which reliable counselor-client predicate matching is assumed to enhance communication. Alternatively, Ein- spruch and Forman referred to Yapko’s (1981) study as “the most sensible research project in the current literature” (p. 592), with findings that PRS matching for hypnotic induction was more effective in the client’s preferred modality than in a nonpreferred modality. Prior to hypnotic induction, however, Yapko had determined the PRS of clients by using a verbal cuing process similar to that used by Hammer– identification by classification of the predominant sensory modality used in verbalizations. In the same section of their article, Einspruch and Forman quoted data from Allen (1982), who used NLP versus massed systematic desensitization (MSD) for the treatment of snake phobia, concluding, “The NLP-based treatment was just as effective as the MSD treatment” (p. 592). What was not reported by Einspruch and Forman was Allen’s finding that neither treatment was significantly different from a control (no-treatment) condition.

Finally, the lack of adequate training of therapists was criticized as a reason to ignore the results of another 11 studies. Although it is accepted that graduate students do not represent highly experienced and senior therapists, the use of such persons is not only traditional in the comparative psychotherapy literature but can also be a strong argument for both the ready application of a specific procedure and the robustness of that procedure. In fact, a great many of the evaluative studies that have been performed on new therapies have been in laboratory situations with just such therapists as those used in the NLP research. These therapists may be seen as the yardstick for comparing a particular therapy with other therapies, and one need only cursorily read the literature on other therapies reported in the journals to discover many successful applications of therapeutic procedures by graduate students. If it is the case that NLP can be demonstrated as effective only by those who have undeigone the “extensive training” that Einspruch and Forman (1985, p. 594) referred to as necessary for effective use of NLP, then NLP may well be a successful (if elusive) procedure. On the other hand, it may be an example of “E bias” in the evaluation of a specific psychotherapeutic environment, in which case it may be the conviction level of the counselor and not the specific treatment or approach to counseling used by the counselor that is the effective variable. As noted by Einspruch and Forman (1985), those practitioners who have “a wealth of clinical data indicating that [NLP] is highly effective” (p. 590) need to provide this data for the wider professional public if the previous nonsupportive conclusions (Sharpley, 1984) are to be challenged at all.

Further Research

The experimental data presented in the 15 studies reviewed earlier was summarized in tabular form (Sharpley, 1984, pp. 240-243) and showed that of those 15 studies of the effectiveness of NLP principles and procedures, 2 were supportive, 5 were uncertain or had mixed results, and 8 included data not supportive of the principle of NLP that they investigated. Examination of the other 22 studies that were identified by Einspruch and Forman (1985) showed that the breakdown was 3 supportive, 4 uncertain, and 15 nonsupportive. Furthermore, although in 1 of the 3 that did support NLP (Brockman, 1980/1981), counselors who used predicate matching were perceived as more empathic than counselors who did not use predicate matching, this finding was contradicted by a number of other researchers. These included Cody (1983), who reported that counselors who matched their clients’ predicates were perceived as less trustworthy than counselors who made no special effort to match predicates. Green (1979/1981) and Dorn (1983a) found that predicate- matching counselors were not perceived as more trustworthy or socially attractive than other counselors, but Appel (1983) noted that opposite sex was more powerful than predicate matching in influencing client preferences for a particular counselor. Atwater (1983) and Ehrmantraut (1983) both found NLP-trained counselors to be no better than “Car- kuff-” or “general-«trained counselors, and Haynie (1982/ 1983) reported that NLP procedures were less effective than no-NLP procedures in developing human relations skills. In another study in which supportive findings were reported (Radosta, 1982), eye movements were used to identify clients’ PRS, but this finding was contradicted by Dorn (1983b), Lange (1980/1981), Talone (1983), Cole-Hitchcock (1980), Fromme and Daniell (1984), and Johannsen (1982). Finally, the outcome reviewed earlier from Yapko’s (1981) study on the effectiveness of PRS matching for hypnotic induction should be considered in the light of Kraft’s (1982) results wherein PRS matching had no beneficial effects on the process of relaxation training.

Seven studies were not reviewed previously by either Ein- spruch and Forman (1985) or me. Of these, Shobin (1980) found that counselor predicate matching increased rapport with clients. Alternatively, Elich, Thompson, and Miller (1985) and Dorn, Atwater, Jereb, and Russell (1983) found no significant data to support the NLP suggestion that eye movements and spoken predicates are indicative of a PRS. Graunke and Roberts (1985) and Ridings (1986) similarly noted no significant support for the accurate identification of a stable and “overriding preference for one sensory modality” (Ridings, 1986, p. 529). Rebstock (1980) found no significant relation between predicate matching and positive evaluation of counselors by clients, but partial support came from Schmedlen (1981), who concluded that predicate matching positively effected clients’ evaluations of counselors on one of the three measures used in that study.

Overview of Research Studies

It was suggested earlier in this article that the number of studies evaluating NLP to date (at least 44) ought to give some indication of support for or rejection of the effectiveness of NLP predicate identification and matching. Of the 44 experimental reports, in only 6 were outcomes reported that vindicate the “uncritical acceptance of NLP, PRS, eye movements, and predicate-matching” referred to by Elich et al. (1985, p. 625). As pointed out by many authors and well stated by Dorn (1983b), “In order for counselors to respond effectively to clients in their PRS, it is important that the PRS be accurately assessed” (p. 154). E>ata collected in 44 studies clearly indicate an overwhelming finding that (a) the PRS cannot be reliably assessed; (b) when it is assessed, the PRS is not consistent over time; therefore, (c) it is not even certain that the PRS exists; and (d) matching clients’ or other persons’ PRS does not appear to assist counselors reliably in any clearly demonstrated manner.

Finally, the sources of the 44 reports should be considered. Table 1 shows that the majority (65.9%) of the studies were reported as graduate theses in Dissertation Abstracts International, with a ratio of nonsupportive (of NLP) results to supportive results of 4.25:1. Studies reported in the journals showed a nonsupportive/supportive ratio of 5:1, with the overall ratio at 4.5:1 against NLP, a fairly consistent finding across the two sources, even considering the 11 studies (mostly dissertations) showing partially supportive or mixed data. It is very difficult to accept (as Einspruch & Forman, 1985, suggested) that all these researchers were guilty of the “methodological errors” (p. 590) that they claimed leave the total research on NLP to date inconclusive and “trivial” (p. 594), Criticism of research on methodological grounds is valuable only when the grounds are well established in the experimental literature (e.g., Campbell & Stanley, 1966; Cook Sc Campbell, 1979). Einspruch and Forman disputed the results of 39 studies on the basis that those researchers did not understand the principles of NLP. Although the present article does not constitute the sort of in-depth summary analysis of outcomes

Table 1

Source and Outcomes of NLP Research Data

Source Supportive Nonsupportive Partial
No. % No. % No. %
DAI (29) 4 13.8 17 58.6 8 27.6
Journals (15} 2 13.3 10 66.6 3 20
Total (44) 6 13.6 27 61.4 11 25

Note. DAI = Dissertation Abstracts International.

recommended by Glass, McGaw, and Smith (1981), it appears that with 44 experimental reports on the effectiveness of NLP, a conclusion can indeed be drawn regardless of the degree to which the various researchers were experts on NLP. The basic tenets of NLP lave failed to be reliably verified in almost 86% of the controlled studies, and it is difficult to accept that none of these 38 studies (i.e„ those with nonsupportive, partial, or mixed results) were performed by persons with a satisfactory understanding of NLP (or at least enough of an understanding to perform the various procedures that were evaluated). If it is true that there are data in the clinical files of proponents of NLP that support it in a way different from the experimental data reviewed, then these need to be published and examined according to the traditional methodological yardsticks of experimental and evaluative literature. Until that time, the inquirer in this field may be forgiven for accepting the conclusion of Elich et aL (1985), “NLP has achieved something akin to a cult status when it may be nothing more than another psychological fad” (p. 625).

Value of NLP: Implications From Research Findings for Counseling Practice

In the opening section of this article, it was suggested that the outcome data on NLP might show (a) a lack of conclusive effects or (b) the essential untestableness of NLP. Became only 13.6% of the 44 studies reviewed supported NLP, one may exclude the first alternative. There are conclusive data from the research on NLP, and the conclusion is that the principles and procedures suggested by NLP have failed to be supported by those data. On the other hand, Einspruch and Forman (1985) implied that NLP is far more complex than presumed by researchers, and thus, die data are not true evaluations of NLP. Perhaps this is so, and perhaps NLP principles are not amenable to research evaluation. This does not necessarily reduce NLP to worthlessness for counseling practice. Rather, it puts NLP in the same category as psychoanalysis, that is, with principles not easily demonstrated in laboratory settings but, nevertheless, strongly supported by clinicians in the field. Not every therapy has to undergo the rigorous testing that is characteristic of the more behavioral approaches to counseling to be of use to the therapeutic community, but failure to produce data that support a particular theory from controlled studies does relegate that theory to questionable status in terms of professional accountability.

What is it then that NLP can offer the practitioner? First, the process of predicate matching to enhance support is worthwhile, and there is a great deal of literature on counselor mirroring that supports the practice of counselors’ using verbal (and nonverbal) behaviors similar to their clients’. This procedure induces empathy, and although this is of great value for effective counseling, this can hardly be claimed as a discovery first made in NLP. Second, assisting clients in moving from one sensory modality to another (e.g., visual- auditory-kinesthetic) to aid in understanding an issue has long been used by Gestalt therapists; this is not an NLP invention either. Further, the process of reframing, or “positive asset search,» has been noted in at least five major therapies besides NLP (Ivey, Ivey, & Simek-Downing, in press) and was most clearly presented by Frankl (1962) in focusing on the positive during his period as a prisoner in a Nazi concentration camp, (For none of these three procedures is the existence of the PRS a prerequisite for application. Even predicate matching itself can be very effectively accomplished by using ongoing counselor responses to client statements [e.g„ Hammer, 1983] without invoking the concept of the PRS or its identification). There are other procedures that NLP suggests are beneficial to counseling (e.g., anchoring, changing history) but that are not in any way specific to NLP. Rather, they may be gleaned from a wide reading of the many counseling theories that abound. As such, 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 (1975, 1976, 1979) posited to present NLP as a new and magical theory. That is, although the proponents of NLP claim its underlying principles (e.g., existence of the PRS, methods of identifying the PRS, and predicate matching as a necessary condition for effective therapy) to be true, they have little to support them and much to answer to in the research literature. If, however, NLP is presented as a “theory-less» set of procedures gathered from many other approaches to counseling, then it may serve a reference role for therapists who wish to supplement their counseling practice by what may be novel techniques for them.

One may conclude that there is little use to the field of counseling research in further replications of previous studies of the principles underlying NLP. In 44 studies of these principles, they have been shown to be without general support from the data. Future research that can contribute new data on this issue via methodological advances or consideration of different aspects of NLP may be justified, but perhaps of more relevance (and value) now would be a careful metaanalysis of the large amount of data already gathered. Efich et al. (1985) referred to NLP as a psychological fad, and they may well have been correct. Certainly research data do not support the rather extreme claims that proponents of NLP have made as to the validity of its principles or the novelty of its procedures.








Power Therapies and possible threats to the science of psychology and psychiatry

Grant J. Devilly

Objective: Advocates of new therapies frequently make bold claims regarding therapeutic effectiveness, particularly in response to disorders which have been traditionally treatment- refractory. This paper reviews a collection of new therapies collectively self-termed ‘The Power Therapies’, outlining their proposed procedures and the evidence for and against their use. These therapies are then put to the test for pseudoscientific practice.

Method: Therapies were included which self-describe themselves as ‘Power Therapies’. Published work searches were conducted on each therapy using Medline and Psychlnfo databases for randomized controlled trials assessing their efficacy, except for the case of Eye Movement Desensitization and Reprocessing (EMDR). Eye Movement Desensitization and Reprocessing has more randomized controlled studies conducted on its efficacy than any other treatment for trauma and thus, previous meta-analyses were evaluated. Results and conclusions: It is concluded that these new therapies have offered no new scientifically valid theories of action, show only non-specific efficacy, show no evidence that they offer substantive improvements to extant psychiatric care, yet display many characteristics consistent with pseudoscience.

Key words: pseudoscience, PTSD, social influence, trauma, treatment.

Australian and New Zealand Journal of Psychiatry 2005; 39:437-445


In 1975 Bandler and Grinder [1] published the first of their two volumes on Neuro Linguistic Programming (NLP). Their book was aptly entitled The Structure of Magic and in it they outlined a revolutionary new method for assessing, communicating with and treating patients. The basic premise was that people are influenced by internal ‘maps’ of information which they gather and organize visually, aurally or kinaesthetically. It was claimed that the trained consultant could identify the method in which the information was stored by eye-gaze patterns, posture, tone of voice and language patterns. It was further claimed that this knowledge facilitated communication during therapy to effect change (e.g. a kinaesthetic representational system would be more amenable to change through the use of ‘feeling’ words during therapy). At the time of its introduction it was heralded as a breakthrough in therapy and advertisements for training workshops, videos and books soon began to appear in trade magazines. The workshops provided certification as an NLP practitioner, advanced workshops led to the title of Master Practitioner and one could even be certified as an NLP trainer. However, by the late 1980s a host of controlled trials had 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 [2].

I refer to NLP here not to target the practice for further denigration, but to hold it up as an early example of what some call an ‘Alphabet Therapy’ and others refer to as a ‘Power Therapy’. To emphasize the issue of fads in psychotherapy what I aim to show is a cycle of business behaviour in our profession. Indeed, one practice within NLP is a technique called ‘Visual-Kinaesthetic Dissociation’ (VKD) and this has subsequently become one of the ‘Power Therapies’ in its own right.

The Power Therapies: healers, sham or spin-doctors?

The term ‘Power Therapies’ arose from the coming together of various therapy advocates on the Traumatic- Stress Forum email list founded by Professor Charles Figley in March 1994. The declared aims of the Traumatic-Stress Forum are: ‘Our most immediate concern is seeking the most powerful, painless and efficient method for eliminating or at least containing the unwanted consequences of traumatic events’ [3]. However, in chasing these worthy goals, advocates of new therapies made unusual and unsubstantiated claims of 100% success rates and one-session-cures using their particular brand of therapy. Each of the therapies had slightly different methods. Some used tapping of ‘energy meridians’, some made people move their eyes backwards and forwards, some relied upon biofeedback-assisted ‘dianetics’ and some were ‘so powerful’ that even a small description of the protocol was prohibited without attending their rather expensive workshops. In testament to evolutionary processes, with each passing month the Traumatic-Stress Forum claims became more outrageous in a effort to overshadow the opposition, until a general truce was called under the shared banner of ‘The Power Therapies’. It is claimed that these ‘Power Therapies’ are at the cutting- edge of psychiatry and psychology and that they are so termed because of their efficiency and efficacy being superior to traditional treatments. Sceptics refer to these treatments as the ‘Alphabet Therapies’, arguing that their major commonality is the use of acronyms and outlandish and unsubstantiated claims.

By 2004 the leading lights under the Power Therapy banner were Eye Movement Desensitization and Reprocessing (EMDR; [4]), Thought Field Therapy (TFT; [5]), Emotion Freedom Techniques (EFT; [6]), Traumatic Incident Reduction (TIR; [7,8]), the Tapas Acupressure Technique (TAT; [9]) and, of course, VKD [10]. A brief summary of each strategy is given here: the origins of the therapy, their claims and a very brief sum-up regarding the evidence for their use. I will leave EMDR to the last because a more in-depth review of this technique (which has more research studies evaluating it than any other technique in the treatment of trauma) is instructive regarding the scientific status and business practices of all the Power/Alphabet Therapies.

Thought Field Therapy [5]

Originally called the ‘5-minute Phobia Cure’ and sold via infomercial advertising, this technique relies on tapping various energy meridians in a specific order for each problem (called ‘algorithms’), while imagining the feared stimulus. It is claimed that this procedure realigns the body’s ‘control system for the disturbing emotions’ ([11], p.l 154). Under this system there are 14 meridians/vessels which lead to approximately 87 billion possible tapping routine combinations. The originator of the technique claims that heart rate variability (HRV: the degree of fluctuation between heart beats) is a valid measure of psychotherapy and has presented various data in an effort to support his claims of TFT efficacy [12]. These data were presented in a special issue of the Journal of Clinical Psychology where the editor allowed seven unreviewed papers on TFT to be published next to seven critiques of the articles. The special issue makes for an interesting, and at times alarming, read. In effect, the evidence forwarded for TFT was predominantly anecdotal and the scientific method usually underlying experimental investigation was frequently misunderstood [13-15]. In reviewing Callahan’s evidence in this special issue McNally [16] stated that he was reminded of Tertullian’s motto: Credo quia absurdum est (I believe because it is absurd). Training in TFT (‘Algorithm Level Training’, ‘Diagnostic Level Training Steps A, В and C’, ‘Advanced TFT Training’ and ‘Voice Technology Training’) can cost over US$300000 with just the ‘Voice Technology’ 3-day training course costing US$100000. There has never been a randomized controlled trial of TFT against an extant, effective intervention.

Emotion Freedom Techniques [17]

Emotion Freedom Techniques grew out of TFT. The founder, Mr Gary Craig (an engineer by trade), apparently ‘spent US$110000 to learn (TFT) at its highest levels’ from Dr Callahan (Personal Communication, Gary Craig, 22 January 1999). He claims that, consolidating the various TFT ‘algorithms’ into a single general purpose, tapping routine (algorithm), EFT is less cumbersome and more efficient than TFT. Indeed, in the same correspondence, Craig claims:

the EFT comprehensive algorithm is also effective 80- 90% of the time for the average client population. In the hands of an experienced practitioner, it goes well into the 90%s. I often (not always) get 90-100% applying it to entire audiences and I don’t even know the specifics of the participants’ various issues. It works just as well even if the tapping is done in ‘reverse order’ from that suggested in the EFT Course. Furthermore, I often do this using only a shortened version of EFT which includes The Setup (psychological reversal correction) and seven tapping points (EB, SE, UE, UN, CH, CB, UA). It takes about

15 or 20 seconds per round. (Personal communication, Gary Craig, 22 January 1999)

There have been two randomized controlled trials into EFT. In the first trial [18], participants who had symptoms matching the DSM-IV criteria for a specific phobia of small animals, received one 30-minute treatment session of either diaphragmatic breathing or EFT. Results displayed a significant treatment effect in favour of EFT. At the follow-up treatment, gains had dissipated to a large extent, although the gains were still improved compared to pretreatment scores on self-report measures and a behavioural avoidance test. In the second trial [19], 122 undergraduate students with self-reported specific phobias were randomly assigned to one of four groups: EFT; placebo (tapping the arms — away from the hypothesized meridians); modelling (tapping a doll instead of themselves); and a control condition (no treatment). Self- reported fear levels (assessed at pre- and post-treatment) displayed no significant differences between the EFT group, the placebo group and the modelling group, which all displayed a significant improvement over time. The control condition, however, did not display a significant decrease in fear ratings. The authors concluded that the apparent gains are, therefore, likely to be because of other non-specific factors such as systematic desensitization, distraction techniques and demand characteristics — the same hypotheses forwarded in explanation for the effects of EMDR [20], another of the Power Therapies. For now, all we can assume is that EFT compares favourably to no treatment or a treatment known to be ineffective for the target presentation. However, for the time being, the benefits appear to be non-specific to the tapping algorithm and unrelated to any putative energy meridians.

Traumatic Incident Reduction [7]

Traumatic Incident Reduction is a direct conversion from Scientology and, in particular, dianetic auditing — a process of discovering old (back to when the participant was 12 months old) ‘repressed’ and painful ‘chains of en- grams’ (memories?) which are then removed to achieve happiness. The TIR process ‘involves repeated viewing of a traumatic memory under conditions designed to enhance safety and minimize distractions’ [21]. There has only been one fully randomized controlled trial. Valentine and Smith [22] compared 56 treated, female inmates who ‘reported’ experiencing trauma to 67 female inmates who were randomly assigned to a waiting list control. Those in the experimental treatment condition were given an unspecified number of treatment sessions in 3–4 hour blocks. Results displayed a ‘trend’ for TIR to be superior to the no-treatment control condition at post-treatment, to produce statistically significant reductions on multiple measures of symptomatology and to build upon these gains to a 3-month follow-up to the point where there were statistically significant differences between the conditions. Interestingly, the control condition also displayed statistically significant reductions on multiple domains over time with a small to moderate effect size (Cohen’s d = 0.34). This study has multiple design problems (e.g. unknown method of randomization, unknown number of treatment sessions, a lack of diagnostic certainty at intake and a lack of information on, and control for, other factors that may affect inmates in correctional facilities). In sum, with the control condition being ‘no treatment’ all we can say for now is that TIR appears to be better than no intervention and unproven against extant, effective interventions.

Tapas Acupressure Technique [9]

Invented by Ms Tapas Flemming (a Californian licensed acupuncturist) in 1993, TAT has been marketed as ‘an easy process for ending traumatic stress, reducing allergic reactions and freeing yourself of negative beliefs’ [23]. The underlying theory is that trauma (and, for some reason, allergies) lead to a blockage of energy in various organs and that applying light pressure to one of four areas (inner corner of either eye, between eyebrows, or back of head) this energy is released and the trauma resolved. Throughout this process the person concentrates on the objectionable material and is also taken through a series of statements. The advocates for this intervention also argue that many allergies are trauma-related and they can be relieved through this process. There does not appear to have been a study of any kind into this technique which has been published in the peer-reviewed literature.

Visual-Kinaesthetic Dissociation [10]

Visual-Kinaesthetic Dissociation is a process whereby the patient imagines the trauma as if watching a videotape of the event from different perspectives, coupled with a temporary dissociation from the event, followed by directed re-association of beliefs regarding the event. It is claimed that the desired dissociation is different from traumatic dissociation in that the desired goal is ‘a shift in one’s perception of a memory from associated (i.e. as if one is reliving the experience) to disassociated (i.e. not experiencing the memory in an associated manner)’ [24]. There has never been a published, peer-reviewed, trial into this technique.

Eye Movement Desensitization and Reprocessing [4]

Shapiro first introduced EMDR with claims of a near 100% success rate for any trauma-related memory within a single session of EMDR [4] and caught the imagination of clinicians and researchers to a possible unitary cure. Until this point posttraumatic stress disorder (PTSD) had been viewed, together with obsessive compulsive disorder, as one of the anxiety disorders most resistant to either psychological or pharmacological therapy. Not surprisingly, Shapiro’s claims led to an overwhelming interest into the technique and led to over 30 dials of the intervention having been completed by 2004. Shapiro now describes EMDR as an eight-phase treatment protocol (see [25] for an analysis of EMDR mutation over time) and labels it as an ‘information processing’ therapy [26]. These eight phases now include: a history taking phase; a preparation phase (e.g. providing coping skills); an assessment phase (e.g. identifying the most vivid visual image related to the traumatic memory, assessing negative beliefs and alternative positive beliefs, obtaining anxiety ratings); a desensitization phase (e.g. following therapists’ fingers with eyes while imagining trauma scene); an installation phase (i.e. challenge negative beliefs with positive beliefs); a targeting of tension phase (with eye movements); a closure phase (homework assigned and feelings normalized); and a re-evaluation phase (techniques taught in phases 1-8 are reassessed for patient learning).

Reviews and meta-analyses of EMDR studies published in high-ranking peer-reviewed journals have consistently found that: there is overwhelming evidence that eye movements are neither a necessary nor a useful addition to the procedure [25,27]; there is strong and consistent evidence that EMDR is better than no treatment and better than ineffective treatments, yet only as good as any other treatment that uses some aspect of exposure therapy [25,28]; and there is growing evidence that a full, exposure-based, intervention package is superior to EMDR in the long term [29,30]. Previous studies not specifically investigating EMDR have shown that distraction techniques used during exposure frequently lead to a dissipation of therapeutic gains and that such dissolution specifically impacts anxiety and depressive symptomatology and becomes more pronounced over time [31-33]. It is argued that such a pattern is beginning to emerge in the case of EMDR [20]. In sum, reminiscent of the charges laid at Jungian psychology, it has been claimed that ‘what is effective in EMDR is not new and what is new is not effective’ ([28], p.619).

A major problem that has plagued EMDR researchers regards the mercurial nature of the protocol itself. In the original paper [4], it was claimed that reading the manuscript was adequate ‘to achieve complete desensitization of 75-80% of any individually treated trauma- related memory in a single 50 minute session’ (p.221), and that further details could be obtained from the author. Enquiries for these details returned a flyer for available workshops — not the more detailed protocol, as is usual in academia. In short order, this declaration of efficacy had changed to ‘while successful treatment without proper training may be achieved perhaps 50% of the time, in other cases, untrained clinicians place the client at risk’ ([34], p.188). By 1994 it was claimed that Level II training was required to properly evaluate the procedure and by 1999 much research was discounted because the treatment adherence raters used in various studies were not sanctioned by the EMDR Institute [35]. By 2004, to become a certified therapist in EMDR, requirements included: EMDR Level II training; at least 2 years experience in the relevant field (e.g. as a psychiatrist or psychologist); having conducted at least 50 EMDR sessions with no less than 25 clients; having received 20 hours of consultation by an approved consultant in EMDR; be recommended for certification by one or more EMDR International Association (EMDRIA)- approved consultants in EMDR; obtaining two letters of recommendation regarding one’s professional usage of EMDR in practice, ethics in practice and professional character; and having completed at least 12 hours of EMDRIA-approved continuing education credits [36]. Every 2 years, 12 continuing education credits in EMDR are also required to maintain certification.

Furthermore, EMDR was initially differentiated from other treatments by the eye movements. As research appeared which showed little utility of inducing eye movements, the protocol insidiously changed. By 1991 it had been changed to include hand or finger tapping instead of eye movements [34], and by 1996 this had extended to ‘any external stimulation’ [37]. It was fortuitous for EMDRIA that researchers were testing the technique by having control conditions which used exactly these behaviours, just before each change in what counted as an acceptable EMDR protocol. When no differences were found between tapping and eye movement conditions, for example, the ground had already been laid to claim that the researchers were simply comparing two forms of EMDR against each other.

Further mutation of the EMDR protocols occurred over time until the ubiquitous one-session cure required at least 5-12 sessions (dependent upon client and trauma type), and included in vivo exposure (an already validated strategy for avoidant behaviours), guided self-imagery and mastery (referred to as ‘positive future templating’ [38]) and, of course, imaginal exposure and cognitive challenging.

Good theories and treatment models do evolve over time. However, such theories explicitly state the conditions under which they could be falsified. Changes over time to the assumptions and procedures should also be made explicit and differentiated from earlier versions to preclude confusion. Failure to meet these criteria results in practices based upon unfalsifiable theories and the utility of spending scientific resources on treatment evaluation becomes questionable.

Commonalities in practice (or how to spot a Power Therapy)

Perusing the above therapies it should become apparent that there are a few factors which unite their protocols. It is known that strategies such as exposure (imaginal and in vivo) and cognitive challenging are effective in treating trauma [39] and all of the above strategies include aspects of these interventions. However, each technique adds its own differentiating twist and domain-specific jargon, whether they be eye movements, body tapping, ‘voice technology’, ‘installation phases’ or targeting of ‘chainedengrams’. It is, therefore, unsurprising that these strategies tend to prove better than no treatment or nonspecific treatments. Indeed, misattribution of treatment efficacy to accoutrement practices might lead me, for example, to hypothesize that it is the green leather chair that I use during imaginal exposure for PTSD treatment which is the active ingredient to my therapy and responsible for symptom relief. Does this give me the right to advertise my new Green Chair Therapy? It would be interesting to see the medical fraternity’s reaction to a ‘new, revolutionary treatment’ for bacterial infections that consisted of prescribing the cutting-edge ‘bluer-than-blue’ coloured amoxicillin. With this in mind a popular satirical website is that of a Dr Fatima Shekel, which markets a ‘revolutionary breakthrough in trauma treatment’ called Sudotherapay [40]. Selling a Sudometer to aid in treatment and offering Sudodoctorates (SyD) and certification in Sudotherapay up to Level IX. This approach is even conducting an online ‘clinical’ trial.

But how did these interventions obtain such a widespread following of practitioners? It is my thesis that there are certain other commonalities amongst the Power Therapies that are best examined through a social psychology lens. It is claimed that social influence strategies are commonly used by those peddling pseudoscience. Pratkanis [41] identifies nine tactics that are frequently used (sometimes unknowingly) by pseudoscientists and, using the Power Therapies as examples, they are as follows.

First, Pratkanis [41] recommends the creation of a phantom — a currently unavailable goal that might just be obtained with the right angle, effort or insight. For example, EMDR proponents originally claimed a one-session cure for any trauma-related memory. This process is naturally aided by the suspension of disbelief and frequently relies on one’s hopes that ‘maybe it is possible — let’s have a closer look’. Second, it is suggested that the aspiring pseudoscientist set a rationalization trap. With this tactic the pseudoscientist encourages further commitment to the technique by obtaining incremental commitments to the protocol. For example, trainings might start-off with a free ‘information’ session (or pamphlet) on the technique which is quickly followed by an invite to be trained to Level 1 at the ‘knock-down’ price of only $233. After all, if you have spent a night listening about the technique (or reading about it) you must be interested, surely? Following this a larger ‘investment’ is suggested (e.g. for Level 2 training) and so on. In effect, the target (e.g. psychiatrist) rationalizes that they must be interested as they have already invested substantial time and money into the practice. It is also no accident that these trainings are held at plush, five star hotels which convey a sense of credibility whilst at the same time pairing a positive affect with the technique.

Third, Pratkanis suggests that the purveyor of pseudoscience manufactures source credibility and sincerity. This involves the creation of a guru-like leader with special (or specialized) traits. It is indeed quite difficult to argue against someone of stature — particularly one trained to Level 9 in a ‘Power Therapy’ and who understands ‘voice technology’. But it is even harder to argue with someone who is seen as ‘gifted’ and affects ostentatious compassion towards those in strife. Maybe they set-up a ‘humanitarian’ (and tax exempt) offshoot, such as the EMDR Humanitarian Assistance Program, or maybe all they do is sign all correspondence with the word ‘hugs ’ instead of ‘yours sincerely’, as in the case of the founder of EFT. But whatever the method, faked credibility coupled with implied sincerity is a mighty force to oppose.

Pratkanis suggests that the budding entrepreneur then creates a ‘granfalloon’. This term was first introduced by Kurt Vonnegut [42] in the phrase ‘if you wish to study a granfalloon, just remove the skin of a toy balloon’ to describe what he later defined as a ‘proud and meaningless association of human beings’ [43]. All the Power Therapies promote at least one self-regulated body of followers who have in-group behaviours, rituals, jargon, shared goals and feelings and specialized information. For example, when EMDR was first introduced, training workshop delegates were required to sign a seemingly legal binding document stating that they would not train others in the technique and would not show therapists untrained in EMDR the treatment manual that was distributed during the training. Aside from ensuring a lack of competition for the trainers, such a tactic declares those trained in the process as somehow ‘ special ’ and facilitates in-group thinking.

These in-groups are then more likely to rally to the call against enemies (out-group critics). It is particularly important to have enemies (i.e. critics), for without them there is little in the way of publicity, and few scapegoats, should none of your manuscripts be accepted by peer- reviewed journals (which raises another important issue, namely ‘what do we mean by peers’). This was not lost to P.T. Bamum, self-acclaimed as ‘the greatest showman on earth’ (and frequently misattributed with the saying ‘there’s a sucker bom every minute’), who started his illustrious career with a travelling circus of ‘freaks and oddities ’ and quickly learned that any publicity was better than none. Whenever he moved to a new town he would write a letter of complaint (under a pseudonym) to the local paper the morning after the circus had opened. He even complained that many of the acts were cons and advised the readership to go nowhere near the circus. Of course, people thronged to his shows which became famous in very short periods of time, and it is unsurprising that later in life his most popular lecture was ‘the art of moneygetting’. Such social influences, for example, have led the Journal of Clinical Psychology to allow a special issue of unreviewed articles on TFT to be published, next to critiques of the articles, following claims by TFT advocates that ivory-tower pedagogues would never allow their breakthrough to see the light of published day — so threatened would they be by such results from unorthodox techniques. Of course, TFT is now marketed with the tagline ‘published in the Journal of Clinical Psychology’ [44], adding yet more source credibility to their claims.

The use of self-generated, persuasion is thought to be one of the most powerful social influences [41]. Here, in true Amway fashion, the customer becomes the seller. Should someone be coaxed or coopted into selling the product the incurred cognitive dissonance [45] increases the belief in the product and acts to increase or maintain a social identity consistent with the group (i.e. groupthink; see [46]). These customers/sellers may then resort to vivid appeals to argue their case [21]. Being concerned with mental health, therapists tend to have a natural inclination towards compassion for human suffering. With respect to this the single case description very much humanizes the problem and coopts the listener to not disagree with the message in order to continue to be seen as a compassionate human being — particularly if the case study being relayed is by the actual patient. It makes it very difficult for the decision maker (e.g. psychiatrist or case manager) to appear to be a caring individual while publicly disbelieving the advocate or turning down approval for wholesale delivery of therapy X, having been told an anecdotal case study. Furthermore, humans tend to have a confirmatory bias [47]. We always remember the horse that won the race, but not the multitude which lost, the time we put it all on number 4 and the chips came-up. But, of course, we remember the case where treatment X worked and not the times the clients never returned or did not improve. Consequent to such a bias, the neophyte Power Therapists may well believe their own vivid appeal to others, yet for scientists, case studies generate hypotheses and pique interest, but they also understand, as the adage would have it, that the plural of anecdote is not facts.

As mentioned earlier, TFT advocates have argued that heart rate variability is a valid measure of technique efficacy and have advanced such data in ‘proof’ of the protocol. The budding pseudoscientist may then argue that the burden of proof is on the critic — not those doing the claiming, as is usual in science. This is just one example of prepersuasion, where the advocate sets the stage for what should be counted as evidence for their product. One might even create whole new disorders that your therapy is then alone in being shown (via case studies, of course) to have curative powers (e.g. ‘love pain’ being ‘cured’ through EFT [48]). Another example of this tactic is to set expectations for a certain outcome which, because of the above-mentioned confirmation bias, then becomes a self-fulfilling prophecy (e.g. therapist allegiance effects in EMDR trials [20]). In sum, prepersuasion tries to stack the deck in favour of the dealer.

Pratkanis [41] goes on to suggest that taking advantage of various human heuristics (simple rules humans frequently use to govern decision-making) will increase the allure of the product. These heuristics include: the scarcity heuristic (if it’s scarce or costs more, it has more intrinsic value); the consensus or bandwagon heuristic (if key people or most people agree, then it must be true); the message length heuristic (the longer, the stronger); and the representative heuristic (if the protocol is complex it will be good for treating complex cases). Likewise, he argues that commonplaces (widely held beliefs in today’s society) also act to increase the perceived value of the product. These commonplaces include: the natural commonplace (natural energy intervention is good, man-made medicine is bad); the goddess within commonplace (people have a spiritual self which is neglected by science and modern medicine); and the science commonplace (having ‘scientific’-sounding aspects to the product adds to its credibility).

Finally, when all else fails, Pratkanis recommends that one should attack critics with innuendo and ad hominem arguments. Playing the man rather than the ball becomes far more profitable when one does not know the rules of the game, one is inept at the game, or where one is obviously losing. This can take quite a few forms (e.g. implications of a flawed character through not being compassionate, charges of a general lack of therapeutic ability to explain poor outcomes with the new treatment, insinuations of being protective over the scientific status quo etc.) but the most subtle and insidious form in recent years has been the implied threat to sue any negative press (e.g. published randomized controlled outcome trials, commentaries etc.) as defaming a copyright name. Were this not the case, it is unlikely I would have had a high ranking scientific journal send my article outlining a randomized controlled trial to a lawyer before publishing it, and I doubt very much that McNally’s article comparing EMDR to Mesmerism would normally contain the front page note T am very grateful to those who have reviewed previous drafts of this article including Harvard University attorney Frank J. Connors, J.D., attorney Kathleen Moore, J.D., Margaret Dale, J.D.’ ([49], p.225). It seems that pointing out that the emperor has no clothes is becoming more fraught with liability as the years pass.

A more recent tactic has even been the discounting of the entire scientific method. Power Therapists have, in some cases, adopted the (quite illogical) postmodern mantra that science subjugates personal meaning. An example of this is the dismissal of science as unnecessary because of personal accounts and anecdotal case studies, as mentioned earlier. Another is a direct attack on both the null hypothesis and the application of critical ex- amination/thought. In an online forum where there was some discussion regarding a lack of correlation between the claims frequently made by Power/Alphabet Therapists and the experimental data, one advocate unwittingly showed this tactic rather concisely:

Feeling rather tired of the scientism rampant in our professions, of the people who would rather devote their time to trying to disprove things than to helping people directly, and not feeling much patience right now — maybe I need to tap on that, (open correspondence to the Traumatic Stress Forum, 18 February 1999)

It is too easy to attribute these types of comments to one or two misguided individuals, but it is becoming a rather frequent response type in psychology and psychiatry. There is a difference between understanding the limitations of, and generalistic conclusion brought about through, randomized controlled trials and meta-analyses and the overall discounting of the scientific method.

Rucker and Pratkanis [50] argue that ad hominem insults, or accusations in general, are very powerful methods of influencing opinion when one accuses the target of negative traits that one is actually guilty of oneself. In a series of experimental studies, they found that such ‘projection’ was effective even when the audience had evidence that the claimants were, in fact, guilty of this deed themselves. In a sad indictment of the ease with which humans are erroneously persuaded, Rucker and Pratkanis also found that the audience still laid blame on the target and became more sympathetic to the one doing the accusing even after suspicions had been raised about the motives of the projectionist.

A caveat for psychology and psychiatry

There is a rather important caveat that needs to be borne in mind when inspecting Pratkanis’ methods ‘to sell a pseudoscience’ when considering psychotherapies: many, if not all, empirically supported psychotherapies meet at least some of these criteria! For example, nearly all have a charismatic leader, have established organizations devoted to their use and use vivid appeals to proliferate their use. The major difference, however, is that empirically supported practices build upon a scientific theory and state the terms under which this theory could be falsified. In effect, all scientific theories are tentative. I contend that this is the most important condition which delimits a science from a pseudoscience [51]. As summarized by Karl Popper in 1963:

Thus the problem which I tried to solve by proposing the criterion of falsifiability was neither a problem of meaningfulness or significance, nor a problem of truth or acceptability. It was the problem of drawing a line (as well as this can be done) between the statements, or systems of statements, of the empirical sciences, and all other statements — whether they are of a rehgious or of a metaphysical character, or simply pseudo-scientific. Years later — it must have been in 1928 or 1929 -1 called this first problem of mine the ‘ problem of demarcation’. The criterion of falsifiability is a solution to this problem of demarcation, for it says that statements or systems of statements, in order to be ranked as scientific, must be capable of conflicting with possible, or conceivable, observations. ([52], pp.38-39)

Another differentiating factor is the declaration of a coherent and consistent theory that underlies the practice — one which, besides being testable, does not disagree with currently understood and accepted experimental data. The therapies outlined above have all been derived in isolation of scientific theory and in many cases the theory has been subsequent to the practice (e.g. EMDR). Shallowness of theory is not a good enough reason in itself to dispel practice, but it should raise concern and suggest that such practices should be approached with caution. As commented by Lewin [53]: ‘Many psychologists working in an applied field are keenly aware of the need for close cooperation between theoretical and applied psychology. This can be accomplished in psychology, as it has in physics, if the theorist does not look towards apphed problems with high-brow aversion or with fear of social problems, and if the applied psychologist realizes that there is nothing so practical as a good theory’ ([53], p.169).


Community mental health is becoming a larger focus of governments and organizations as the years progress and rightly so. This has led to greater spending and more emphasis on evidence-based practice. However, without a grounding in what ‘evidence-based practice’ actually means, this has resulted in a larger market for pseudoscience. Practitioners, bureaucrats and the general public do not always understand the tenets of science and, as such, are more open to being duped. This is not to say that investigating new avenues and new hypotheses are a waste of resources, but rather that we need to set, and abide by, stringent conditions as to what counts as evidence and the methods we use to evaluate hypotheses. However, possibly the largest hurdle in this field facing psychology and psychiatry is the education of our practitioners and the wider community into the scientific method and internal and external threats to its integrity. As referred to above, McNally resorted to Tertullian’s motto (I believe because it is absurd), yet the clever hawker of pseudoscience would be well aware of St Augustine’s variation: credo ut intelligam (I believe so that I might understand).

I see two major steps that are needed immediately to help counter this proliferation of pseudoscientific practices within the field of mental health. The first involves the education of extant mental health professionals in the tenets of science and the tactics of pseudoscience. This will be best met by continuing education focusing on this larger issue and promoting the work of organizations such as the Commission For Scientific Medicine and Mental Health which pubhshes journals such as the Scientific Review of Mental Health Practice and The Scientific Review of Alternative Medicine. Consequently, this also raises the thorny issue of not crediting certain workshops as meeting the minimum requirements needed for continuing education credit. Second, it is imperative to train new students in the scientific method and alert them to practices which threaten the scientific integrity of the profession. Courses have been proposed, and are indeed now run, based upon differentiating science and pseudoscience within the mental health professions (e.g. see the edited book by Lilienfeld et al. [54]).

To end where I began — NLP is no longer as prevalent as it was in the 1970s and 1980s, but is still practised in small pockets of the human resource community today. The science has come and gone yet the belief still remains. In fact, you can enrol in an Australian workshop today for certification as: an NLP Practitioner ($3995); a Master NLP Practitioner ($4395); or an NLP Trainer ($10 570). The companies offering the training will even arrange finance. Be quick, places are limited!












Limited evidence that neurolinguistic programming improves health-related outcomes


Question: What are the effects of neurolinguistic programming (NLP) on health-related outcomes?

Outcomes: Studies were eligible if they had examined any health-related outcome in any population. Identified studies reported on 18 outcomes; the majority of studies measured anxiety (n=6) and three studies measured quality of life and depression. Other outcomes measured included weight maintenance, morning sickness, substance misuse and claustrophobia during MRI. Eleven of the 18 outcome measures were self-reported, 3 were objective (weight, successful completion of MRI scan, urinalysis for illegal substances), 2 were observed and 2 were not reported. Time-of-outcome assessment ranged from immediately post-treatment to up to 3 years.


Design: Systematic review of experimental studies (randomised controlled trials (RCTs) and pre-post test studies). Data sources: MEDLINE, PsychINFO, ASSIA, AMED, CINAHL, Web of Knowledge, CENTRAL up to February 2012. Additional searches included NLP specialist databases at the Universities of Bielefeld and Surrey, the European Association for NLP, other NLP associations, research groups and social network forums. Hand-searching of reference lists was also performed.

Study selection and analysis: Primary research studies were eligible if they examined the effects of NLP on any health-related outcomes in any clinical population and reported results quantitatively. Studies examining single NLP techniques were excluded. Risk of bias was assessed for the pre studies and poststudies using the Downs and Black quality index score.


Ten studies met inclusion criteria, nine of which were published in peer-review journals and one identified online. Five studies were RCTs and five were pre-post studies. Six studies reported participant age (range from 7 to 75 years), eight studies reported participant gender (2 studies conducted in women only; the remainder 64% female), and one study reported sociodemographic data (100% with higher education). NLP delivery, described in six studies, was carried out by clinical psychologists, psychotherapists and certified NLP practitioners. Sessions were 1-2 h in length, with three studies using a single session and the remaining studies providing between 4 and 20 sessions, delivered over between 1 and 12 months. The main outcomes assessed by the RCTs were anxiety (self-report), child development (observed), weight (objective) and quality of life (self- report), comparing NLP with no intervention or with active control. In four of the five RCTs, NLP had no significant effect on any of the outcomes assessed. In one RCI) NLP reduced psychological distress and increased quality of life compared to waiting list control. Three RCTs and five pre-post studies reported within group improvements from baseline. Of the three studies measuring objective outcomes, one reported a post-treatment increase in completed MRI scans and two had no post-treatment improvements in urinalysis for illegal substances or weight maintenance.


There is currently little evidence to demonstrate that NLP interventions improve health-related outcomes. Further high-quality studies are needed.


No studies reported results by intention to treat analysis. Risk of bias was high in three RCTs because of alternate group allocation and incomplete outcome data reporting. Only 1 of the 10 studies (an RCT) had a blinded outcome assessor.















University of Glamorgan


Neuro-linguistic programming (NLP) is a popular form of inter-personal skill and communication training. Originating in the 1970s, the technique made specific claims about the ways in which individuals processed the world about them, and quickly established itself, not only as an aid to communication, but as a form of psychotherapy in its own right. Today, NLP is big business with large numbers of training courses, personal development programmes, therapeutic and educational interventions purporting to be based on the principles of NLP.This paper explores what NLP is, the evidence for it, and issues related to its use. It concludes that after three decades, there is still no credible theoretical basis for NLP, researchers having failed to establish any evidence for its efficacy that is not anecdotal.

Key words: Neuro-linguistic programming, theoretical credibility, cargo-cult psychology.



Neuro-linguistic programming (NLP) is a school of thought founded on the psycho-therapeutic ideas of Richard Bandler and John Grinder. Since the publication of their co-authored book, The Structure of Magic in 1975 (in which Bandler and Grinder describe NLP as therapeutic magic), NLP has developed into a world-wide phenomenon. A simple Google UK search reveals a plethora of organisations and individuals offering NLP for training, personal development, coaching, and as an intervention aid for eating disorders, addictions, dyslexia, depression and chronic fatigue syndrome, to name but a few. NLP has been described by Tosey and Mathison (2003) as: «… one of the world’s most popular forms of interpersonal skill and communication training» and is a recognised form of psychotherapy according to the United Kingdom Council for Psychotherapy.

To the casual observer, NLP appears to be a widely accepted set of techniques. Indeed, NLP has found its way into a number of academic institutions, appearing in peer-reviewed journals from an array of disciplines including counselling, business, marketing and education. This gives the impression that is not only widely used but is academically credible with a sound research base to support it. In short, NLP presents as a technique that we should all be aware of. It presents as though its central ideas should be universally available since it represents a model of human behaviour that can dramatically improve communication skills, empathy, and indeed, troublesome thought processes. Despite the cloak of respectability, the truth about NLP borders on the worrying. This paper argues that NLP is an ill-defined chameleon that masquerades as a discipline open to the rigours of academic enquiry, when in fact there is spectacularly no evidence to support NLP beyond personal testimony and anecdote.

What is NLP?

The term neuro-linguistic programming conjures up an air of scientific respectability, yet its very name is wholly inappropriate. O’Connor and Seymour (cited in Skinner and Croft,

2009) explain why this particular nomenclature was used:

  • ‘neuro’: refers to our neurology, our thinking patterns.
  • ‘linguistic’: language, how we use it, and how we are influenced by it.
  • ‘programming’: refers to the patterns of our behaviour and the goals we set.

Bandler is reported to have stated that «neuro-linguistic processing» was a term that he made up to avoid having to be specialised in one field (Skinner and Stephens, 2003). This would constitute a forgivable admission were it not for the persistence of its use today, and the pseudo-scientific, yet totally misleading, connotations of the term.

Firstly, our thinking patterns should be defined as ‘cognition’ not ‘neuro’. Use of the latter word is effectively fraudulent since NLP offers no explanation at a neuronal level and it could be argued that its use fallaciously feeds into the notion of scientific credibility. ‘Linguistic’ again makes associations with the academically credible field of linguistics. And how does ‘programming’ equate to the patterns of our behaviour and the goals we set – aren’t these ‘behaviours’ and ‘thought processes’? Indeed, ‘programming’ actually implies a lack of conscious thought processes.

The links with scientific credibility persist in NLP books: «NLP is the art and science of excellence» (O’Connor and Seymour, 1994, cited in Heap, 2008). Yet despite this, and despite its very name suggesting strong links with accepted science, NLP has no credible basis in neuroscience and has been largely disowned by the very academic fields within which it claims to lie, namely psychology and linguistics.

What are NLP’s central ideas?

NLP was founded on central philosophies born out of Bandler and Grinder’s observation of transcripts and films of psychotherapy sessions. In particular, Bandler and Grinder were influenced by the hypnotherapist, Milton Erickson; the family therapist, Virginia Satir; and the founder of Gestalt Therapy, Fritz Peris. They considered these therapists to have a reputation for success and sensibly wanted to attempt to learn from their techniques. However, as Heap (2008) points out, what resulted was not a set of techniques based on good practice, but rather a number of suggestions of the ways in which we behave, think and communicate.

A core principle proposed in NLP is the notion of a preferred representational system (PRS). It is suggested that individuals construct internal maps of the world by processing external information through five sensory systems: visual, auditory, kinaesthetic, olfactory and gustatory. It should be noted that in the context of NLP ‘kinaesthetic’ inexplicably refers to feelings in general. It is suggested within NLP that a person’s conscious activity predominantly uses one of these systems (particularly visual, auditory and kinaesthetic) and, according to Grinder and Bandler (1976), the particular system being used at any given time is reflected in that individual’s style of speaking.

An individual thinking in the visual mode, for example, will tend to predicate sentences with visually-related words such as: «I can see that…» or: «It looks to me as if…». Bandler and

Grinder (1979) also claimed that the representational system an individual uses at any given time can be revealed in their eye-movements. For example, it is proposed that the kinaesthetic mode is associated with a downward gaze to the right. Given that Grinder and Bandler (1976) proposed that each individual has a preferred idiosyncratic representational system, it follows that two individuals perceiving the world through different systems will be having differing experiences of that world. In order to achieve maximally effective communication, NLP proposes the notion of matching, whereby one individual matching the verbal and non-verbal behaviours of another individual can tune into their representational system and hence, to their view of the world.

What is the evidence for NLP’s central ideas?

If the claims of Bandler and Grinder were substantiated, then it would be true to say that they had uncovered a corner stone of human cognition. They are claims that easily lend themselves to empirical investigation and, in the 30 years since the claims were first made, volumes of supportive research evidence should be available to underpin these theories being taught in university psychology departments across the world. Three decades on, however, the most striking observation about the perpetuation of NLP is that it exists almost entirely in isolation from published evidence to substantiate it. The core ideas of NLP from the mid 1970s were mostly discredited in the 1980s. Sharpley (1984) reviewed the research to date concerning NLPs assertion of a PRS and concluded that that there was little evidence for the use of a PRS in NLP, with much data to the contrary.

Even prior to NLP, mainstream psychology had been investigating the link between hemispheric asymmetry (reviewed by Ehrlichman and Weinberger, 1978) and eye movements, so it was not unreasonable for Bandler and Grinder to propose a link. However, in terms of the specific claims made by NLP, the supportive evidence is scant and at best offers only partial support. Wertheim eta! (1986), for example, examined the hypothesis that eye-movements reflect sensory processing. Consistent with Bandler and Grinder’s claims, Wertheim and co-workers found evidence of increased upward eyepositioning and stares when participants were asked to recall visual information but findings from the auditory and kinaesthetic modalities were inconsistent. Further, Wertheim and colleagues (1986) dismissed any notion of their findings being supportive of NLP since auditory-type eye position changes were most prevalent in all three (auditory, visual and kinaesthetic) stimulus conditions. Beyond this one study, evidence for Bandler and Grinder’s claim is notable by its absence from the cognitive psychology literature. Surely this must be because cognitive psychology tested the claims and failed to find an effect.

In response to criticisms, Sharpley (1987) updated his earlier review with further evidence reporting that of 44 studies evaluating NLP, only six could be categorised as accepting the principles of NLP, PRS, eye movements, and predicate-matching without criticism. Sharpley quantified the credibility gap further by pointing out that the majority of studies were not published in peer-reviewed journals but appeared to be abstracts from postgraduate theses. The ratio of non-sup- portive to supportive studies was 4.5:1, and Sharpley concluded:

(a) the PRS cannot be reliably assessed;

(b) when it is assessed, the PRS is inconsistent over time; therefore,

(c) it is not even certain that PRS exists; and

(d) matching clients’ or other persons’ PRS does not appear to assist counsellors reliably in any clearly demonstrated manner

Sharpley (1987, p105)

The lack of a credible research base is not unknown by the NLP community. Consider the following quote from the University of Surrey’s NLP research project website:

The academic research into NLP is thin. The empirical studies to date have various limitations (we review this research in a forthcoming journal article).

We believe there is an urgent need for more research, of a variety of methodological types. It is sometimes believed that the only valid research and the only type in which academics are interested, is experimental and uses statistical methods to develop proofs. This is a narrow and somewhat stereotyped view of research. We support, in particular, quantitative [sic] and action- based methods, and we are strongly interested in the potential of NLP ‘modelling’ as a phenomenological research method.

In addition to pursuing our own research, seeks to support academic researchers and NLP practitioners wishing to enquire into NLP and its applications.

Neuro-linguistic Programming and Research (2006) Centre for Management Learning and Development

University of Surrey

Phenomenological research is free from hypotheses, pre-conceptions and assumptions, and seeks to describe rather than explain. Given the claims made by proponents of NLP, this adds little to the credibility debate and would produce reports concerning the experience from the perspective of the individual rather than confirmation of the claimed efficacy.

The fact remains that NLP proponents make specific claims about how NLP works and what it can do and this compels providing evidence to substantiate these claims. The above statement constitutes an admission that NLP does not have an evidence base and that NLP practitioners are seeking a post-hoc credibility.

Can NLP be though of as an umbrella term?

Criticisms of the primary ideas of NLP have more latterly been addressed with the argument that NLP has evolved to encompass the modelling of effective strategies in top performers and the adoption of strategies in others towards achieving a desired outcome. Craft (2001) argues that NLP draws on the theoretical framework of social constructivism – thus it is considered to be experiential, action-based and involving the negotiation of meaning. Tosey and Mathison (2003), while concurring with Craft (2001) that NLP is a set of strategies rather than a theory, suggested it was possible to infer a theoretical cohesion and that NLP should be described as reflecting a systemic theory drawing its inspiration from the work of the cyberneticist Gregory Bateson. As such, NLP can be considered to be focused on feedback mechanisms

As Linder-Pelz and Hall (2007) state, NLP is about adopting a humanistic constructivist approach involving collaboration, focus on solutions, precision questioning, detachment from the problem, feedback and finding out what works and what doesn’t. However, such a description appears to categorise NLP as anything that ultimately helps an individual address a particular life issue. There exists in this an evaluative problem. An individual meets with an NLP practitioner regarding a particular issue. Strategies are tried until ultimately the individual feels a solution has been found. The practitioner thus claims another success story. Within this however, it is impossible to quantify precisely what has happened owing to the humanistic constructivist label. In this context, to describe NLP as social and/or humanistic constructivism is nothing more than tautology and creates a smoke screen around the conclusion that its core ideas are unsupported.

The use of NLP as an umbrella term only adds to the confusion and conveniently excuses its proponents from having to substantiate its claims.

«At the end of the day none of this matters because NLP really works» – or does it?

If NLP encourages people to learn ways of communicating more effectively then that is a noble endeavour and not particularly problematic. However the problem arises with the perpetuation of claims. It has been suggested that NLP is: «being applied widely, if often informally, in UK education» (Tosey and Mathison, 2003, p371). Such informal application makes it difficult to assess, but the claims of one NLP website are fairly typical, claiming that NLP can help you:

  1. Discover the children’s preferred learning styles and allow for them to be different.
  2. Use circle time to share their values and identity.
  3. Celebrate their sunbeams and re frame their raindrops.
  4. Allow children to share how they do things so that they can model each other.
  5. Use brain gym to calm, energise or reconnect right and left brain for improved concentration.
  6. Help the children to access an appropriate state to learn easily.
  7. Increase motivation by recognising success and putting it in the future.

New Oceans (2005)

Brain Gym® (referred to in claim 5) is a commercial learning efficiency programme that appears to have been taken up by some schools, despite a complete lack of evidence for its efficacy (Hyatt, 2007) and is beyond the considerations of this paper. Of the remaining claims: 2, 3 and 7 are simply shallow statements with 1, 4 and 6 based on NLPs discredited claims about learning styles. In short, these claims are simply nonsense.

In addition to the potential for informal application in education, ‘NLP-certified practitioners’ make claims about its efficacy in the treatment of a whole range of quite serious disorders such as addictions, eating disorders, anxiety problems and pain management to name but a few (Brain-train, 2007, for example), yet the medical literature is devoid of any published evidence to substantiate these claims. This creates a serious ethical problem in both the educational and the paramedical fields. As Heap (2008) points out, knowledge is power and anybody making claims about being able to help with serious disorders or improve learning efficiency is making a claim for some kind of power. However, with that power, there must be accountability through public scrutiny. The lack of evidence for such claims means that the most rudimentary test of accountability cannot be addressed. In addition to this, if NLP is just a communication model, what special abilities does obtaining a certification in it bestow upon an individual which allows them to meddle in education issues and serious medical conditions?

In relation to dealing with vulnerable (indeed perhaps desperate) people, the claims of unqualified practitioners are extremely worrying. The precise nature of a ‘qualification’ in NLP is difficult to ascertain with many organisations offering impressive sounding training from ‘Diplomas’ up to ‘Master Practitioner’. Precisely who accredits these ‘qualifications’ though? Who is responsible for externally examining and moderating them? How are they regulated? And how long do they take? The latter point is key with training courses in NLP being offered over a period of as little as two days. Consider the training required to become a Chartered Clinical Psychologist – a British Psychological Society- (BPS-) accredited first degree is needed, followed by three years of doctoral-level training within the National Health Service. Entry to the doctoral courses is fiercely competitive and so successful applicants have usually worked as psychological assistants for a number of years. The whole process is regulated by the BPS (NB in 2009 the Health Professions Council will become the regulator), who in addition to setting the framework for ethical practise, have a discipline and complaints procedure that is crucially administered by independent non-psychologists. Such a system ensures that individuals are not only appropriately qualified, but are publicly accountable for their actions. Similar training is required to specialise in the other professional areas of psychology (health psychology, educational psychology, counselling psychology, forensic psychology, occupational psychology and sports psychology) with a minimum of six years training. An individual presenting themselves as being a ‘Master Practitioner’ in NLP is giving the impression of having acquired a high level of training, yet it is an unregulated ‘discipline’. A code of conduct has been set out by the Association for Neuro-linguistic Processing, yet worryingly it contains the following disclaimer:

The Code does not assume that individual Members possess particular levels of skill in any specific area; it is important, therefore, that users of Members’ services do satisfy themselves that the person they are working with is appropriately skilled

Association for Neuro-linguistic Processing, 2007

To put the onus of responsibility onto the individual seeking the service is scandalous. What basis do they have to satisfy themselves that an individual is qualified in the face of impressive sounding claims and ‘qualifications’?

Personal testimonies are not difficult to come by in relation to the efficacy of NLP. A Google search will again yield a wealth of personal testimonies and endorsements of the powers of NLP. Given that a similar search will equally yield personal testimony in favour of many other dubious techniques such as homeopathy, astrology or even trepanning, such testimonies are of little worth. Carl Sagan (Sagan and Druyan, 1996) suggested a number of ways of detecting a fallacious argument (now known as ‘Sagan’s Baloney Detection Kit’), the most pertinent being wherever possible there must be independent confirmation of the facts. Such independent confirmation of the claims of NLP does not exist.


One could argue that to refute NLP is to engage in argu- mentum ad ignorantiam. However, NLP singularly fails to stand up to scrutiny concerning its face validity and its construct validity. NLP’s predictive validity is more difficult to ascertain as proponents of the ‘discipline’ engage in academic goal-post shifting and arguments about its ‘constructivist’ nature. Claims about what NLP can do persist though and as such it is analogous to Bertrand Russell’s celestial teapot with the burden of proof to support its theoretical foundations and efficacy as an intervention lying with its proponents.

The physicist Richard Feynman coined the term ‘cargo cult science’ (Feynman, 1985). In the South Seas there is a cargo cult of people who, during war-time, observed lots of airplanes carrying goods. They wanted the planes to continue to land after the war ended and so set about reconstructing airports with fires alongside the runway, a wooden hut for the air traffic controller to sit in and antennas made of bamboo. Despite the form of the airport being right, the planes didn’t land! Feynman adapted the idiom of ‘cargo cult science’ to refer to research that follows all the form and pretence of scientific investigation yet is missing something essential.

To adapt this term one more time, NLP masquerades as a legitimate form of psychotherapy, makes unsubstantiated claims about how humans think and behave, purports to encourage research in a vain attempt to gain credibility, yet fails to provide evidence that it actually works. Neuro-linguistic programming is cargo cult psychology.








Thirty-Five Years of Research on Neuro-Linguistic Programming. NLP Research Data Base. State of the Art or Pseudoscientific Decoration?

The huge popularity of Neuro-Linguistic Programming (NLP) therapies and training has not been accompanied by knowledge of the empirical underpinnings ofthe concept. The article presents the concept of NLP in the light ofempirical research in the Neuro-Linguistic Programming Research Data Base. From among 315 articles the author selected 63 studies published in journals from the Master Journal List of ISI. Out of 33 studies, 18.2% show results supporting the tenets of NLP, 54.5% — results поп-supportive of the NLP tenets and 27.3% brings uncertain results. The qualitative analysis indicates the greater weight of the non-supportive studies and their greater methodological worth against the ones supporting the tenets. Results contradict the claim of an empirical basis of NLP.

Keywords: neuro-linguistic programming, NLP, pseudo-science, psychotherapy



For more than twenty-five years therapies, personal development training, courses and other forms of working with people advertised as based within the Neuro-Linguistic Programming (NLP) Framework have enjoyed enormous popularity on the market of psychological services. NLP practitioners are found among university employees, and advertisements of NLP-related institutions appear in popular science magazines. Students of psychology attend courses where they attain successive degrees of initiation for NLP practitioners. NLP trainings have been provided in companies such as Hewlett-Packard, IBM, McDonald’s, NASA, the U.S. Army, and U.S. Olympic teams, and in countless public school systems (Singer & Lalich, 1996). It has been suggested that NLP is “being applied widely, if often informally in UK education” (Tosey & Mathison, 2003, p. 371). I investigated official psychology curricula of the 12 best state universities in Poland. Eight offered contents and, in many cases, even separate courses devoted to NLP (Witkowski, 2009).

And still, despite that widespread presence of NLP, none of the psychology textbooks that I have heard of (Polish or English) presents an in-depth discussion of the concept. What’s more, scientific authorities refrain from giving their opinions in this respect. What is the scientific worth of the concept? Only a thorough analysis of empirical research can lead to the answer.

Outline of the NLP concept

In the 1970s, Richard W. Bandler and John Grinder came up with a brilliant idea to create a practical therapy model. They argued that outstanding psychotherapists acted on the basis of implicit theories, which ensure their effectiveness and great rapport with patients. Furthermore, they concluded that observation of the most skilful therapists, their contemporaries, at work should result in the discovery of patterns, which could be then generalized, verified on an empirical basis and put into therapeutic practice. For several years they observed such therapists as Fritz Peris, Milton H. Erickson and Virginia Satir at their work. The gathered material enabled them to formulate NLP tenets and hypotheses.

The central philosophy of NLP is summed up in the sentence “The Map is not the Territory” (see, e.g. Lankton, 1980, p. 7). That means that each of us operates on the basis of our internal representation of the world (the “map”) and not the world itself (the “territory”). The maps that we create are mostly limited and distorted. The therapist’s task is to understand and operate on the basis of the client’s map of the territory.

The maps that people make of their world are represented by five senses: visual; kinesthetic, referring totactical and visceral sensations; auditory, including noises and sounds; olfactory, including smell; and gustatory,including taste. Each experience in the world of senses is composed of information received through the said systems of

senses, different in terms of quality, which are termed representational systems by the NLP original proponents (Grinder & Bandler, 1976; Bandler & Grinder, 1979). They suggested that each of us processes the majority of information using one primary representational system (PRS). Following the example of the most outstanding therapists, to work effectively with a patient one should necessarily match the patient’s PRS so as to be able to use their “map”.

Another discovery of which the NLP originators were particularly proud of was to realize that access to the representational systems is possible through the so-called accessing cues that are precisely specified eye movements. Careful observation of these movements should enable the NLP therapist to unequivocally identify the PRS of the patient, interlocutor etc., and, in consequence, facilitate matching their PRS. All other hypotheses of the NLP system related to the arising of mental disorders, the type of therapy and communication, etc. stem from these basic assertions.

When analyzing how the NLP concept was formulated, it is worth indicating analogies between the manner in which it had been developed and the research methodology applied in social psychology proposed and defined by Cialdini (1980) as a full-cycle approach to social psychology. Bandler and Grinder followed the full-cycle method, but regrettably they omitted the stage of empirical verification of their assertions. They found that part of the process inessential and moved straight to the formulation of the system and putting it into practice. Bandler, known for his openly demonstrated contempt for scientific testing of the NLP hypotheses, claimed that his system represented an art, not science, hence testing its assertions was pointless or even impossible. The NLP founders distorted the full-cycle approach creating a quasi-cycle process, which included only these three stages. Against the contempt expressed by Bandler, the NLP system being used so widely made many researchers test its theoretical underpinnings on an empirical basis.


Selection of Material for Analysis

In order to obtain a coherent empirical image of NLP, independent from beliefs of therapists and subjective opinions of academic psychologists, I conducted an analysis of the majority of scientific articles devoted to NLP ever published. A most extensive register of such studies termed the Neuro-Linguistic Programming Research Data Base (State of the Art) is to be found on the web pages of NLP Community ( rdb.cgi). At present it supplies abstracts and bibliographic information with reference to 315 articles, by and large empirical, written by 287 authors and published in the years 1974-2009. The base was created at the University of Bielefeld in Germany in 1992, and moved to Berlin in the later years. It was designed to gather and organize empirical available studies concerning NLP from all over the world. The base is referred to by its creators as “state of the art”, being updated and recommended on an ongoing basis by numerous institutions worldwide, which draw extensively on NLP in their activities. In Poland this base is recommended by e.g. Polski Instytut NLP (The Polish Institute for NLP) whose founder and chairman — Benedykt Peczko — personally suggested it to me as the most all- embracing global source of scientific studies on NLP. Out of several bases of articles developed by NLP proponents this one offers the highest number of entries. The analysis I am reporting in this article was carried out in December 2009.

There were three major arguments in favor of my choice of this database. Firstly, I came to the conclusion that the 18 years of work on the base performed by people committed to showing empirical underpinnings of the concept must give better results than those I could have achieved if searching through other available bases in a short time, such as PsychLit, PsycINFO or MEDLINE. Secondly, the fact of using the base established by followers of the concept might meet their possible accusations that I was biased and partial in preparing my review. Thirdly, analysis of the base contents, of the manner in which it is updated and of selection of articles might disclose additional information on how the image of NLP as a science with empirical foundations is created.

Quantitative analysis

In order to obtain the very essence of the empirical material available in the base, I performed a number of operations on the base. The first was to select the most reliable studies for further analysis. To this end, I evaluated them based on the criterion of whether the journal in which the given articles were publisher was recorded on on the Master Journal List of the Institute for Scientific Information in Philadelphia. This operationdoes not require justification in more detail. Although there are many doubts raised to this list, magazines from the Master Journal List are much less likely to have published unreliable articles than others. As a result of the initial selection, of 315 articles I had 63 — accounting for 20% of the entire base — left for further analysis.

Of interest are the findings of quantitative analysis of publications in individual years. By reference to the diagram including all 315 studies, it is clear that scientific activity peaked in the eighties of the 20th century, (see Figure 1). It experienced a minor renaissance at the beginning of the present century. Based on the diagram, one may assume that as a research issue NLP enjoyed

Only Master Journal List All publications from NLP Research Data Base

Figure 1. Number of all publications included in NLP Research Data Base in individual years as against the number of studies of Master Journal List.

Journal of Multicultural Counseling and Development 1 Journal of Nonverbal Behavior 1

Journal of Social Psychology 1

Management Decision 1

Neuropsychologia 1

Psychosomatics 1

SA Pharmaceutical Journal 1

Total Quality Management and Business Excellence 1

immense popularity in the period directly following the formulation of its empirical underpinnings in the seventies of the 20th century. In the subsequent years the research interest in NLP was decreasing. The bottom diagram shows the Master Journal List publications exclusively. Activity of researchers having their studies published in renowned magazines was proportional to the entire sample.

The sample of 63 studies selected for further analysis included articles published in 30 different magazines. Below is a breakdown of the number of articles published in individual magazines.

Journal of Counseling Psychology 12

Perceptual and Motor Skills 10

Psychological Reports 6

Professional Psychology: Research and Practice 4

Psychological Bulletin 3

American Journal of Clinical Hypnosis 2

International Journal of Clinical and Experimental Hypnosis 2

Psychological Science 2

American Journal of Family Therapy 1

Anaesthesia 1

Assessment and Evaluation in Higher Education 1

Brain and Cognition 1

British Journal of Clinical Psychology 1

European Psychologist 1

Gerontologist 1

International Journal of Hospitality Management 1

International Journal of Language and Communication Disorders 1

Journal of Abnormal Psychology 1

Journal of College Student Development 1

Journal of Consciousness Studies 1

Journal of Consulting and Clinical Psychology 1

Journal of Marital and Family Therapy 1

This high number of magazines may be treated as indirect verification of the reliability of the gathered empirical evidence. It will be difficult to maintain that one of the magazines or a group of them was biased in favor of NLP or that their activities were aimed at deprecating the concept. It should also be emphasized that the thematic scope of the magazines, which published the studies devoted to NLP was very wide indeed.

I put the selected sample of articles through a qualitative analysis, as a result of which three categories of studies emerged:

  1. Thirty-three empirical articles, which tested the tenets of the concept and/or the tenets-derived hypotheses.
  2. Fourteen articles comprising polemics, discussions, case analyses, or empirical works in which NLP represented little significant aspect etc. that is studies of no empirical worth from the point of view of my analysis.
  3. Sixteen works having nothing in common with the NLP concept, available in the base most likely by chance or due to other reasons that were unknown to me.

The first category is a subject of a more detailed quantitative and qualitative analysis presented in the subsequent part of this article.

The second category comprises such studies as, for instance, a phenomenological account of the first author of horse riding lessons (Mathison & Tosey, 2008), a discussion

Table 1

Number of articles published in individual magazines and their weight.

Magazine Cat. 1 Cat. 2 Cat. 3 Weight
Journal of Counseling Psychology 8 3 15
Perceptual and Motor Skills 1 5 2 10
Psychological Reports 3 1 10
American Journal of Clinical Hypnosis 1 15
International Journal of Clinical and Experimental Hypnosis 1 24
Psychological Science 1 24
British Journal of Clinical Psychology 1 24
Journal of Abnormal Psychology 1 24
Journal of Consulting and Clinical Psychology 1 24
Journal of Multicultural Counseling and Development 1 10
Journal of Social Psychology 1 15
Neuropsychologia 1 24
Total Quality Management and Business Excellence 1 10

Note. Cat. — category. Weight — based on the scoring system of the Polish Ministry of Science and Higher Education allowing for the impact factor of individual magazines.





of the application of lateral concepts in the new millennium (Corballis, 2000), an analysis of a new therapy carried out by Virginia Satir (Woods & Martin, 1984), an analysis of 2 cases of rape victims (Koziey & McLeod, 1987), an analysis of a case of recovering from clinical depression (Hossack & Standidge, 1993) and many other (Davis & Davis, 1983; Beck & Beck, 1984; Yapko, 1984; Einsprach & Forman 1985; Dailey, 1989; Peterson-Cooney, 1991; Adler, 1992; Witt, 2003; Brown, 2004). Accepting the analyzed sample as 100%, the articles on NLP, which proved to be useless for the purpose of empirical analysis, represented 22.2%.

I found the third category puzzling. It encompassed, inter alia, an essay on changes in Soviet psychology on the path of perestroika (Gindis, 1992), social and ethical limitations of psychology (Drenth, 1999), intuition (Lieberman, 2000), mimicry (Lakin, Jefiferis, Cheng & Chartrand, 2003; Stel, Dijk & Olivier 2009), status of the pharmacists (White, 2009), application of alternative therapies to children with dyslexia (Bull, 2009) and many others (Norcross & Prochaska, 1983; Malloy, Mitchell & Gordon, 1987; Kamiol, 1995; Starker, Pankratz, 1996; Norcross, Hedges & Prochaska, 2002; Cullen & McLaughlin, 2006; Boden & Giaschi, 2007; Abramowitz & Lichtenberg, 2009; Супа, Andrew & Tan, 2009). Articles not related to the NLP concept account for as much as 25.4%. What is interesting, articles from this category represented an insignificant share in the eighties of the 20th century, which increased gradually towards contemporary times.

Studies published in magazines of the Master Journal List of Institute for Scientific Information in Philadelphia constitute the essence of the empirical material with reference to NLP. There were as many as 33 such papers representing 52.4% of the selected sample. The qualitative analysis allowed me to single out the following subcategories:

  1. Nine works supporting the NLP tenets and the tenetsderived hypotheses (27.3%).
  2. Eighteen works non-supportive of the NLP tenets and the tenets-derived hypotheses (54.5%).
  3. Six works with uncertain outcomes (18.2%).

Sources of individual articles and the number of articles of individual categories published therein are shown in Table 1.

Not to be satisfied with the quantitative indicators of the published articles only, in the last column of Table 1, I present the scoring taken from the list of scientific journals available from the Polish Ministry of Science and Higher Education. The scoring allows for the impact factor of individual journals and constitutes the basis for the assessment of the worth of the scientific output of researchers in Poland. Totaling up points received for individual publications will provide a better illustration of their weight. If it is assumed that all those papers were written by one scientist, in Poland they would have been given 479 points (100%), i.e. 123 points (25.7%) for the NLP supportive articles, 281 points (58.6%) for the studies challenging the tenets of the concept, as well as 75 points (15.7%) for the works showing uncertain outcomes. As it appears, after this conversion the significance of the studies with negative outcomes increased as against the ones with positive results.

The present analysis is my second performed on the database in question. A similar one four years earlier in December 2005 and published the findings in 2006 (Witkowski, 2006). The base has been considerably expanded since then. In 2005 it had 180 studies published in the years 1977-2005. At present the base covers the 1974-2009 period. This does not mean however that more than 100 papers on NLP were written throughout the last four years. The analysis of the base contents indicates that it has been supplemented with an additional one study from 1974, thus increasing the number of works in the years 1977-2005 by 107 entries, and 31 articles published in the years 2006-2009. Some works have also been removed from the base, i.e. one entry from 1977, two from 1987, one from 1994 and one from 1997.

Qualitative analysis

The numbers indicate unequivocally that the NLP concept has not been developed on solid empirical foundations. Less than one-third of the analyzed works shows supportive evidence, more than a half — non- supportive, and the remaining papers — uncertain results and doubts. But let’s move beyond the mere numbers. Argumenta ponderantur, non numerantur — the force of the arguments lies in their weight, not numbers. It is often the case that one study weighs as much as a number of others. Some works verify basic assumptions of a theory, others only a less significant aspect of the problem.

The studies reporting outcomes, which I rated as supportive of the concept tested its basic assumptions in a very small number. In this respect, the study by Kinsboume (1974) is exceptional as it tested the hypotheses concerning eye movements, as well as Yapko’s experiment (1981) revealing that the matching primary representational system had a positive influence on the depth of hypnotic relaxation as compared with the control group. Dooley and Farmer’s study (1988) may also be possibly classified into this category.

A high number of the remaining papers in this category lacked control groups. Most frequently only the initial and final measurements of the same group of subjects were taken. This was the case with the research of Duncan, Konefal and Specifier (1990), who provided 21-day residential training to a group of subjects, and then compared the pre-training and post-training status based on self-actualization measures of the Personal Orientation Inventory. Two years later Konefal, Duncan and Reese (1992) carried out almost identical research. The 21-day training was also provided, and this time the authors measured changes in trait anxiety and locus of control. In the subsequent research performed under a similar procedure and without a control group either, Konefal and Duncan (1989) measured changes in social anxiety. Studies on the application of NLP by employees of Southern India companies were conducted without any control group (Singh & Abraham, 2008), and so were studies on the application of NLP for treating post-traumatic stress disorder (Muss, 1991). The latter work from this category showed positive effects of neuro-linguistic mirroring in cross-cultural counseling (Sandhu, Reeves & Portes, 1993).

It is most likely that any type of intensive 21-day effort undertaken on self-development, based on any concept, would result in similar changes as those measured in the quoted research. The placebo effect is relatively frequent both in therapy, as well as in other forms of social influence

Similarly, it is difficult to state whether positive effects of mirroring resulted from application of a specific NLP technique or from the necessity to put more focus on observation of the interlocutor, which in turn, was positively evaluated by them.

With respect to the category of non-supportive articles, the majority of studies concerned the basic NLP tenets. Several works were devoted only to tests of the eye movement hypothesis (Thomason, Arbuckle & Cady, 1980; Farmer, Rooney & Cunningham, 1985; Poffel & Cross, 1985; Burke et al., 2003). They all provided unequivocally negative results. The preferred modality was researched into by Gumm, Walker and Day (1982), and also by Coe and Scharcoff (1985). In both cases the results did not support the nemo-linguistic programming theory.

Other studies tested NLP tenets in a more complex manner, investigating several hypotheses in one study. Fromme and Daniell (1984) researched into the imagery and sensory mode, as well as communication. They were unable to find any support for the NLP-derived hypothesis that subjects showing differential ability across sensory modes would choose word phrases reflecting their preferred sensory mode. No support was found for the NLP-derived hypothesis that subjects matched for visualization ability would communicate information more accurately than would mismatched subjects.

Elich, Thompson and Miller (1985) tested claims that eye movement direction and spoken predicates are indicative of sensory modality of imagery. Again, these tests did not find any support for NLP-derived hypotheses.

Graunke and Roberts (1985) tested the impact of imagery tasks on sensory predicate usage. The findings proved to be incongruent with R. Bandler and J. Grinder’s conceptualization of representational systems.

Particular attention should be given to two reviews of research (Sharpley, 1984; 1987). In the first, the authorcarried out a thorough analysis of 15 other research studies. What is interesting, as many as 11 of these works are not available in the database in question. A few conclusions from that review are worth quoting:

(…) the identification of this PRS (if it is a PRS and not merely current language style) by either eye movements or self-report is not supported by the research data. (…) The existence or stability of the PRS is irrelevant to predicate matching as a counseling process, and parsimony argues for the process rather than the yet unverified theory. (…) Of most importance, there are no data reported to date to show that NLP can help clients change, (p. 247)

The second review (Sharpley, 1987) is even more conclusive. It was written as a response to a critical paper by Einspruch and Forman (1985), in which the authors analyzed 39 studies devoted to NLP indicating methodological errors and a lack of sufficient knowledge about the theoretical underpinnings of NLP demonstrated by authors thereof. Sharpley took into account works analyzed by Einspruch and Forman, expanded that sample with seven additional ones and performed an analysis similar to mine, reviewing 44 studies (of which two are not included in the base either). Six papers (13.6%) provides evidence supportive of NLP-derived theses, 27 (61.4%) failed to lend support for the NLP tenets, and 11 (25%) shows only partial support. The author investigated all available works starting from the doctoral dissertations to those published in high scoring magazines. This is how he summed up his review:

There are conclusive data from research on NLP, and the conclusion is that the principles and procedures suggested by NLP have failed to be supported by those data. (p. 105) Certainly research data do not support the rather extreme claims that proponents of NLP have made as to the validity of its principles or the novelty of its procedures, (p. 106)

The subsequent three studies referred to the influence of counselors’ or therapists predicate matching on the effectiveness of their actions and quality of rapport (Dowd & Pety, 1982; Dowd & Hingst, 1983; Ellickson, 1983). None provided support for NLP-derived predictions.

Studies on the effectiveness of specific therapeutic techniques failed to support the NLP tenets, too. Krugman, Kirsch and Wickless, (1985) tested Bandler and Grinder’s claim for a single-session cure of anxiety. They did not find support for this claim. Similarly, Matthews, Kirsch and Mosher (1985) tested the effectiveness of double hypnotic induction. Comparison of the experimental group against the control group did not support the hypothesis. In addition, application of pacing and metaphor to overcome client resistance did not support the Bandler and Grinder’s claims (Dixon, Parr, Yarbrough, Rathael, 1986).

Additionally NLP proved to be of little use as a method of enhancing human performance considered by the US Army (Swets & Bjork, 1990). “The conclusion was that little if any evidence exists either to support NLP’s assumptions or to indicate that it is effective as a strategy for social influence.” (p. 90)

The third category comprised six studies with uncertain results. Mercier and Johnson in their research (1984) managed to obtain limited support for NLP theory, with much data contrary to the theory. The same was the case with research by Hammer (1983) on matching perceptual predicates. The findings created more doubts than conclusive data as to perceptual predicates. The researchers studying eye movement as an indicator of sensory components in thought Buckner and Mera (1987) found support for the visual and auditory portions of the model, but the kinesthetic portion was not supported. A similar partial support for the hypothesis that eye movements relate to processing imagery was found in the research by Wertheim, Habib and Cumming (1986).

Other research was carried out based on the assumption that the NLP tenets were true and tested. The examination by Durand, Wetzel and Hansen (1989) may serve as an example here with the researchers analyzing the content of written statements, telephone communications and electronic mail messages in terms of the occurrence of sensory predicate by means of computer software. Similar procedures were followed in other research (Wilbur and Wilbur, 1987).


Among the studies classified as NLP supportive, there was none to indicate in unequivocal terms the existence of different representational systems. Similarly, there was no support found for the claim that subjectswere using primarily one predominant representational system in different life situations. Apart from one study (Yapko, 1981) there is no strong evidence that matching the primary representational system brings beneficial effects in communication and therapy. Two studies supporting some claims for eye movements should be replicated in order to treat their outcomes as supportive of for the hypotheses. Moreover, there are no more extensive and comprehensive research reviews. The only one which might be regarded as such, (Einspruch & Forman 1985) constitutes criticism of the available papers and it does not provide any data to support the NLP tenets. The analyzed works show numerous methodological errors and shortcomings, such as the lack of control groups, and only one research hypothesis being tested or one factor measured.

The studies classified into the non-supportive category are marked by a much higher methodological level. The maj ority allowed for the comparison against control groups, provided measurement of a number of variables, and used a higher number of indicators. Among the studies are two articles offering extensive and high quality research of research. Most results of research from this category were replicated.

Comparison of both categories both in terms of quantity and quality unequivocally indicates the predominance of articles that do not lend support for the NLP tenets, with the ratio of non-supportive to supportive of 3:1. When evaluating the whole empirical research output devoted to NLP, one should also consider the file drawer effect (Rosenthal, 1979). According to it, the NLP supportive studies should have a greater chance for publication then those showing lack of support. It may be easily assumed that a part of the studies that did not find any support for the NLP hypotheses was filed away by researchers.

Review of the articles issued in sources other than those from the Master Journal List indicate the existence of many review works showing the lack of any NLP underlying principles as well. Two of them are worth mentioning. Heap (1988) analyzed 63 studies and concluded thatthe assertions of NLP writers concerning representational systems have been objectively and fairly investigated and found to be lacking. In consequence the hypothesis about the possibility to identify PRS through careful observation of eye movements was not confirmed either. In Heap’s view, these conclusions, and the failure of investigators to convincingly demonstrate the alleged benefits of predicate matching seriously question the role of such procedure in counseling. Dom, Brunson, Bradford and Atwater (1983) also concluded from their review of the literature that there was no demonstrably reliable method of assessing the hypothesized PRS.

While conducting my analysis I noted a certain historical aspect of NLP supportive research. As I realized, most of the research was carried out in the 1980s and partially in the 1990s. In the subsequent years, the number of such research studies decreased and they concerned secondary aspects of the concept or were performed based on the assumption that the fundamental principles of NLP are true. The world of science was apparently losing its interest in the concept of Bandler and Grinder, having confronted it with the research findings. The concept’s proponents lacked motivation to undertake any type of research into, for instance, the effectiveness of its methods.

Another facet, which is worth discussion and that emerged during my analysis is the matter of investigating how the data base is utilized by its administrators, as well as its users. The base is commonly invoked by NLP followers and indicated as evidence for the existence of solid empirical grounds of their preferred concept. It is most likely that most of them have never looked through the base. Otherwise, they might have come to the conclusion that it provides evidence to the contrary — for the lack of any empirical underpinnings. Moreover, they not only fail to browse through the database, dare I say, but they also do not read articles available therein. Reading of two review papers (Sharpley, 1984, 1987) would enable them to first discover that the base lacks as many as 13 entries and then to update it. Fortunately for the present analysis, the missing 12 entries are not included on the Master Journal List.

The number of theoretical studies in the base, such as polemics, dissertations, and discussions is so high that referring to it as to the Research Data Base is considerable misinterpretation as well. What is even stranger is the fact that works completely unrelated to NLP are added to the base. While reading such articles I strengthened my belief that it was only due to some single key words that the NLP related status of those papers was approved. This gives rise to the suspicion that even the database administrators do not read articles, not to mention the abstracts.

All of this leaves me with an overwhelming impression that the analyzed base of scientific articles is treated just as theater decoration, being the background for the pseudoscientific farce, which NLP appears to be. Using “scientific” attributes, which is so characteristic of pseudoscience, is manifested also in other aspects of NLP activities. It is primarily revealed in the language — full of borrowings from science or expressions referring to it, devoid of any scientific meaning. It is seen already in the very name — neuro-linguistic programming — which is a cruel deception. At the neuronal level it provides no explanation and it has nothing in common with academic linguistics or programming. Similarly impressive sounding and similarly empty are expressions used for formulation of tenets of the concept, such as sub-modalities, pragmagraphics, surface structure, deep structure, accessing cue, and non- accessingmovement.

My analysis leads undeniably to the statement that NLP represents pseudoscientific rubbish, which should be mothballed forever. One may even come to believe that my analysis was a vain effort after all. It yielded the same conclusions as the ones arrived at by Sharpley (1984, 1987), Heap (1988) and others. Without doubt, NLP represents big business offering and tempting people with amazing changes, personal development and, what is worst, therapy. In this respect the analysis is an update of the state of knowledge on the subject by reviews published in the period after the latest analyses. Furthermore, is also provides arguments sufficient to answer the following ethical question: Is using and selhng something non existent and ineffective ethical?

The response will surely be similar to the statement given once by Einspruch and Forman (1985) — the effectiveness of NLP therapy undertaken in authentic clinical contexts bytrained practitioners has not yet been properly investigated. Additionally we will certainly be told that NLP works and this should be sufficient reason to use it. Nevertheless, the burden of proof with respect to finding evidence of the effectiveness of the NLP therapy hes on proponents, not skeptics. Here I would like to refer to the statement expressed by O’Donohue and Ferguson (2006), who propose that each type of therapy that does not have empirical supportive evidence of its effectiveness should be called experimental. They also put forward a suggestion that each case of performing such therapies without informing the clients about its experimental status should be referred to and treated as criminal activity. I fully agree with this view. We do not even imagine pharmaceutical concerns marketing medicines whose side effects are uncertain or unknown, yet we allow any psychotherapy to be practiced, in many cases — without any relevant research.

If the NLP assertions on the existence of PRS as well as on the possibility to enhance communication just through matching proved to be tme, it would revolutionize neurosciences, cognitive psychology and some other disciplines. If the NLP claims on the instant effectiveness of the proposed therapies proved to be true, the entire area of psychotherapy would turn upside down and research reports with respect to the effectiveness of therapy would have to position the NLP therapy at the top. Nothing like this is taking place. Instead we find NLP on the list of discredited therapies. Norcross, Koocher and Garofalo (2006) sought to establish consensus on discredited psychological treatments and assessments using Delphi methodology. A panel of 101 experts participated in a 2-stage survey, reporting familiarity with 59 treatments and 30 assessment techniques and rating these on a continuum from not at all discredited (1) to certainly discredited (5). Neuro-linguistic Programming for treatment of mental/behavioral disorders averaged 3.87 (SD=0.92).


The analysis of the NLP Research Data Base (state of the art) by all measures was like peeling an onion. To reach its core, first I had to remove some useless layers, and once I arrived, I was close to tears. Today, after 35 years of research devoted to the concept, NLP reminds one more of an unstable house built on the sand rather than an edifice founded on the empirical rock. In 1988 Heap passed a verdict on NLP. As the title of his article indicated, it was an interim one. In the conclusions he wrote:

If it turns out to be the case that these therapeutic procedures are indeed as rapid and powerful as is claimed, no one will rejoice more than the present author. If however these claims fare no betterthanthe ones already investigated then the final verdict on NLP will be a harsh one indeed (p. 276).

I am fully convinced that we have gathered enough evidence to announce this harsh verdict already now.