Cognitive Behaviour Therapy and Positive Deviancy

Henck van Bilsen

Starting from the premise that there is a need to liberate Cognitive Behaviour Therapy, it is argued that using positive deviancy as a method to harness the creative potential of CBT practitioners could be a way of achieving this objective.


Cognitive behaviour therapy (CBT) has travelled a long way from its humble beginnings as the ‘underdog’ therapy to its current position of dominance. No other form of psychotherapy has produced so many evidence-based models that explain variations of human suffering. It is therefore with great sadness that I make the observation that CBT is going back in time. I notice a development from the information-rich, theory driven, idiosyncratic functional analysis to a diagnosis-recipe model. In other words, CBT is retreating from embracing a psychological model and advancing towards a medical model.
As a supervisor and CBT trainer I have been prompted to start thinking about this subject. The following examples from my practice are typical of what students and graduates of a variety of CBT courses regularly bring for discussion.

I need a diagnosis

The supervisee presents with much concern that, after 5 sessions, she still is not completely certain which DSM diagnosis fits the client’s presentation. This is important she states because if she decides on the incorrect diagnosis, she might do irreparable harm to the client. ‘It is as if I would give the wrong medication. Should I follow a social phobia protocol or a GAD protocol or is the diagnosis panic and agoraphobia?’ This budding CBT therapist was paralysed by the presentation of the client as it did not fit easily into a diagnostic category.

I did not bring my forms so had to cancel my session

The supervisee apologises for not being able to present a recording of one of his training cases. He had to see three clients in a local GP surgery, but arrived there without his portfolio of supporting forms (disorder-specific formulations/models; homework form; thinking errors form; thought record; information on specific disorders leaflets) and felt so de-skilled that he cancelled the session. A telling quote: ‘If I can’t show the client the form with the CBT model of their diagnosis, I feel naked. It is as if I know nothing.’

And in session two we started with the exposure and thought records

The supervisee presents, glowing with pride, that he started with exposure and thought record work in session two having diagnosed during the first session that the client had a specific food phobia. Questions about the client posed by the supervisor and fellow students (client’s reasoning process in avoiding certain foods, the emotion involved, client’s life and functioning other than the food problem) were met with silence as the supervisee had been seduced by the quick diagnosis and deemed further inquiry unnecessary.

These are just a few examples of many supervision situations I have encountered in the last couple of years. Students from a range of CBT courses present with questions and problems like this; something has changed. Until a couple of years ago supervision was often focused on finding the idiosyncratic mechanisms that created and maintained clients’ problems and both supervisee and supervisor realised that ‘going slow meant going fast’ (Sun Tzu, 2005).

Cognitive behaviour therapy – from freedom to straightjacket

CBT really started with early attempts to apply learning theory to behaviour change (Watson and Rayner, 1920). Further theorising and laboratory experimentation on learning theory was done by Skinner (1938), Salter (1949) and Mowrer (1950). It took more than three decades, from the first experimental studies on basic learning processes and the formal beginnings of the behavioural component, for what we now know as CBT to emerge.

This development took place on three continents: Africa, Europe and The Americas. Wolpe (1958, 1969) in South Africa, published the first structured treatment protocol: reciprocal inhibition to treat anxiety problems. Eysenck (1963) published accounts of treatment methods such as desensitisation, negative practice and aversion therapy, all methods grounded in learning theory. Eysenck was the first to bring these various interventions together under the name behaviour therapy (Farmer and Chapman, 2008). The application of operant learning principles in therapeutic settings was developed in the USA for children  (Bijou and Baer, 1966) and people with learning disabilities (Lovaas, 1987). The term coined here was behaviour modification. The behavioural wave was further developed by Krasner and Ullman (1975) and Patterson (1969); based on earlier work by Kelly (1955), Beck (1963) and Ellis (1958, 2004); it took until the 1980s for cognitive therapy as we know it today to be firmly established.

During these foundation years a core element of cognitive behaviour therapy was the application of cognitive psychology principles and learning theory principles to the problems of clients. Clark (2004) stresses the importance of the interplay between theories, experimental science and clinical practice to produce clinical innovation.

From ‘first principles’ to protocols

A trend in recent years has been the protocolisation of CBT: specific diagnoses are linked to very prescriptive treatment protocols. Could this be called the medicalisation of CBT? It has been demonstrated (Kinderman, 2005; Kinderman & Tai, 2006) that a psychological model is far superior in explaining problems such as depression and anxiety. Within the field of CBT it has also been argued that a focus on diagnostic classifications is NOT helpful (Harvey, Watkins, Mansell & Shafran, 2004). The development in mainstream cognitive behaviour therapy seems to go in the opposite direction. Books focusing on CBT for specific disorders are published in abundance. CBT protocols for specific and identifiable DSM/ICD classifications are produced in vast numbers. The Improving Access to Psychological Therapies (IAPT) programme in the UK is an example of this, as the prescribed treatments are based on disorder-specific treatment protocols and it is stressed that treatment needs to be preceded by having a DSMIV or ICD classification. (IAPT Mental Health Programme, 2008). The assumption is that the specific disorders are real entities and that clients can be fitted into boxes of diagnostic categories.

The problem with this new development is that it could lead to a discouragement of clinical experimentation and an avoidance of using ‘first principles’ (learning theory and cognitive psychology) to make sense of clients’ problems. Great things happen when clinicians try things out, report, and repeat the process until a new or better intervention is developed that can subsequently be tested with reputable research methods.

Training of cognitive behaviour therapists during the foundation years of CBT was focused on understanding the client by building the idiosyncratic narrative or individualised holistic theory about how the client came to have these problems and why they still persist. In other words, first principles were used and applied in this process. Nowadays it seems that giving a diagnosis has become all-important. It seems that the word diagnosis is increasingly being used from a medical perspective; if we have the ‘diagnosis’ (read classification) we know which protocolised treatment to give.

Unfortunately clients in real life often present us with a mixed bag of problems. The trans-diagnostic approach to mental health (Harvey, Watkins, Mansell & Shafran, 2004) debates the use of diagnostic categories as defined for clinical practice. In a categorical diagnostic model (based on DSM or ICD) a group of people is captured under the umbrella of a diagnosis but there are strong moves to focus more on dimensional models, for instance with respect to personality disorders (Widiger, 1992). Harvey, Watkins, Mansell & Shafran (2004) postulate that working with processes that are influential in many diagnostic categories may be a better approach for clinicians and researchers; they state that a trans-diagnostic process is the preferred model:

  • thinking in diagnostic categories leads to ‘us and them’ thinking – stigmatisation
  • with in excess of 350 disorders described in the DSM manual a classification/diagnostic approach is an impossible task (no sane person is able to have this amount of information in their working memory)
  •  in a diagnostic approach valuable personal and idiosyncratic information is in danger of getting lost
  • it becomes very difficult to deal with co-morbidity in a categorical model.

Problems for CBT

A problem with the diagnosis-recipe approach and protocolisation of CBT is that it assumes that CBT outside the protocols is not evidence-based, not real CBT. The main objection to this is that people are always more complicated than their classifications, which leads to practical problems when a purely linear model is followed. The focus on disorder-specific, protocol-driven CBT may result in a generation of CBT therapists obsessed with identifying specific disorders so they can use one of the existing protocols. This could inhibit the further development of CBT (Sarason, 1979; Abma, 2003; Goldfried, 2010)) as it may stifle innovation and creativity. Structuring an idiosyncratic treatment plan for clients will become alien to the modern CBT therapist but, as we have seen, the freedom to deviate from established paths is essential for the further development of treatment models. If Beck and Ellis (1958) had adhered to the psychodynamic protocols and guidelines, cognitive behaviour therapy might not have happened. Perhaps we can learn something from positive deviance.

What is Positive deviance?

Traditionally, deviance refers to intentional behaviours that depart from accepted norms; deviancy is seen as wrong (Spreitzer & Sonnenshein, 2004). Research on deviance with a purely negative connotation is an unnecessarily narrow area of study that overlooks the positive contributions that deviating from the norm can make. The development of a positive deviance construct provides a conceptual framework for understanding the variation in effectiveness of two clinicians both adhering to the same psychotherapeutic approach. The implicit hypothesis of the protocol and diagnosis-recipe people is that outstanding results are achieved by good adherence to the protocols; positive deviance would postulate that we get outstanding results by deviating from the norm, by making small or larger changes to accepted strategies. The two founding fathers of cognitive therapy, Ellis and Beck, were both trained as psychoanalysts; only through deviancy did they come to ‘discover’ a theory and a practice that had more benefits for their patients than the therapeutic method that they had been trained to apply.

When Beck and Ellis started deviating, their reasons for doing so were at first not theory driven; unhappy with the progress their depressed clients were making while undergoing traditional psychoanalysis, they listened to the content of what their patients were telling them and came up with alternative intervention strategies initially without the support of an explicitly formulated theory. Beck and Ellis posited that what they did made sense: more patients recovered. This is an example of positive deviancy in action.

The positive deviance movement offers an important contribution to the understanding of excellence as demonstrated by certain individuals; it adheres to the notion that in every group there are individuals whose norm-deviating behaviours have enabled them to find better solutions to problems faced by others who encounter similar challenges and barriers and have access to similar resources.

Conclusions: Beck to basics

CBT finds itself in a paradoxical situation. On the one hand there is emphasis on protocols, models based on diagnostic categories and uniformity of delivery; on the other hand some of the best applications of CBT advertise creative developments for which there is as yet no hard evidence. For CBT to be fit for purpose in the future we have to go back to a system that would encourage the next ‘Beck’ to develop innovations and not be afraid of the wrath of colleagues that will follow.

Reduce the focus on protocolisation of CBT in training and supervision

The training and supervision of cognitive behaviour therapists needs to find a better balance between the application of protocols and the use of ‘first principles’ for the understanding and treatment of human misery. Currently students of CBT are taught how to follow a cookbook, but when the ingredients do not exactly match what it says in the book, they get flustered and try a different recipe; they need to learn to create a meal from scratch, to work with the rough ingredients and design idiosyncratic treatment plans.

Training needs to include a more theoretical foundation

It pains me to notice that some CBT training programmes have no modules on learning theory. Many students I supervise have only the most basic training in reinforcement theory. Students of CBT need to be able to apply theoretical principles to their practice and learn how the theories underpinning CBT are responsible for the creation and maintenance of psychological problems. Furthermore they need to learn a lot more about how to use the theories underpinning CBT to get them out of therapeutic difficulties. Too often do I see supervisees using a trial-and-error strategy when certain protocolised interventions don’t work; using cognitive psychology or learning theory to overcome therapeutic difficulties seems to have become a skill of the past. If the execution of the treatment plan encounters roadblocks, the budding CBT-ers should learn how to use theory to think themselves out of trouble.

Fostering Thinking-based/Reflection-based Deviancy

The community of cognitive behaviour therapy can be seen as a large tree with many branches. Nurturing new and innovative ways of doing effective CBT needs to be stimulated; practicing cognitive behaviour therapists need to be encouraged to think beyond protocols and to do so in a methodical manner. Training and supervision are the ideal places to start this process on deviancy.

CBT conferences could also be set up, deviancy sharing symposia whereby clinicians could share and comment on case examples and case studies in which the therapist deviated from the accepted route and was successful in achieving therapeutic goals. There may even be a case for the setting up of a CBT deviancy database where professionals could ‘blog’ about deviancy and others would be able to question the deviancy and/or comment on the reasons why it could lead to results. At Canterbury Christ Church University we are looking into ways in which deviancy might be captured; a start has been made by devising a questionnaire that invites therapists using CBT to report their deviancy (


Henck van Bilsen is a consultant cognitive behavioural therapist who originally trained in the Netherlands as a clinical psychologist. He specialises in complex and long-standing problems and regularly presents on motivational issues at national and international conferences.  Henck is an accredited trainer, supervisor and member of the British Association for Behavioural and Cognitive Psychotherapies (BABCP). He currently divides his time between the CBT Partnership (with clinics in London,Tunbridge Wells and  Hertfordshire) and the Canterbury Christ Church University where he is Director of CBT programmes.


Abma, T.A. (2003) Learning by Telling: storytelling workshops as an organizational learning intervention Management Learning 34(2): 199-210, 221-240
American Psychiatric Association (1993) DSM-IV update (January/February) Washington, D.C.
Beck, A. T. (1963) Thinking and depression Archives of General Psychiatry 9:324-333
Beck A.T, Rush A J, Shaw B F and Emery G (1979) Cognitive Therapy of Depression New York: The Guilford Press
Clark, D. M. (2004) Developing new treatments: on the interplay between theories, experimental science and clinical innovation. Behaviour Research and Therapy 42: 1089-1104
Ellis, A (1958) Rational psychotherapy J Gen Psychol.59(1):35-49
Eysenck H. J. (1963) Behaviour Therapy, Extinction and Relapse in Neurosis The British Journal of Psychiatry 109:12-18
Farmer, R.F. & Chapman, A. (2007) Behavioral Interventions in Cognitive Behavior Therapy: Practical Guidance for Putting Theory into Action Washington, D.C: American Psychological Association,.
Goldfried, M.R. (2010) Results of the survey of clinicians experiences in using empirically supported treatments for panic disorder The Clinical Psychologist 63(4): 2-20
Harvey, A., Watkins, E., Mansell, W., & Shafran, R. (2004) Cognitive Behavioural Processes across Psychological Disorders: a transdiagnostic approach to research and treatment. Oxford: Oxford University Press.
IAPT-Mental Health Programme (2008) IAPT Implementation Plan: National Guidelines for Regional Delivery DH/Mental Health Programme/Improving Access to Psychological Therapies
Kelly, G. (1955) The Psychology of Personal Constructs. Vols I & II New York: Norton,
Kinderman, P. (2005) A psychological model of mental disorder Harvard Review of Psychiatry 13:206–217.
Kinderman, P. & Tai, S. (2006) Clinical implications of a psychological model of mental disorder Behavioural and Cognitive Psychotherapy 35:1–14.
Lovaas O. I. (1987) Behavioral treatment and normal educational and intellectual functioning in young autistic children J Consult Clin Psychol 55(1):3–9
Mowrer O. H. (1950) Learning Theory and Personality Dynamics New York: Ronald Press,.
Patterson, G.R. (1969) Families: A social learning approach to family life Eugene: Castalia Pub. Co.
Sarason, I.G. (1979) Three lacunae of cognitive therapy Cognitive Therapy and Research 3(2): 223-235
Salter, A. (1949) Conditioned Reflex Therapy: the direct approach to the reconstruction of personality Wellness Institute | 2002 (repr.1949)
Skinner B.F. (1938) The Behavior of Organisms: an experimental analysis Cambridge, Massachusetts: B.F. Skinner Foundation
Spreitzer, G. M. & Sonenshein, S. (2004) Toward the Construct Definition of Positive Deviance American Behavioral Scientist 47(6):828-847
Sun Tzu (2005) The Art of War El Paso: Norte Press
Ullman, L. P., & Krasner, L. (1975) A Psychological Approach to Abnormal Behavior NewJersey: Prentice-Hall
Watson, J.B. and Rayner, R. (1920). Conditioned emotional reactions Journal of Experimental Psychology 3(1):1–14.
Widiger, T.A. (1992) Categorical versus Dimensional Classification: implications from and for research Journal of Personality Disorders 6:287-300.
Wolpe, J. (1958) Psychotherapy by Reciprocal Inhibition California: Stanford University Press
Wolpe, J. (1969) The Practice of Behavior Therapy Oxford: Pergamon Press


Image: Modelli e invidiosi by contrasto_gp


Please enter your email address below to receive notifications about Contemporary Psychotherapy:

Something went wrong. Please check your entries and try again.


20 Lonsdale Road

In association with The Minster Centre, London.


Opinions expressed in this journal are solely those of the author(s).
Publication in Contemporary Psychotherapy does not imply endorsement of those views.


Copyright belongs to Contemporary Psychotherapy. Material may only be reproduced with written permission from the Editor. Authors may use their own material elsewhere after publication without permission. Production of single copies for personal use is allowed without special permission.

Scroll to Top