Cognitive Behavioural Therapy (Brilliant Business) 1st Ed.

5. Using behaviour to change your mind

If I drop (a) a normal house brick and (b) one the same size made of paper, from a diving board, which will hit the water first? If you haven’t studied physics recently, the instinctive response for most people is that the house brick, being heavier, will fall faster. In fact this is not the case. The force of gravity accelerates both objects at the same rate and, because air resistance, not mass, makes the crucial difference, both will both hit the water at much the same time. You still don’t believe me? What would convince you? Maybe if you saw it with your own eyes?

According to tradition, Galileo performed just such a live demonstration by dropping a canon ball and a musket ball from the leaning tower of Pisa. Equivalent trials have been successfully repeated many times since and the findings hold. If you still need to see for yourself you can watch the Apollo 15 experiment on YouTube.

The point is that sometimes seeing really is believing, especially if what we are witnessing is at odds with our expectations. Knowledge gained from first-hand experience can affect us much more deeply than things we know logically must be true or that we have heard second-hand, even from the most reliable source.

Cognitive behavioural therapy recognises this, and this is why the approach makes extensive use of behavioural experiments in helping people address their negative thoughts and feelings.

Often when people complete thought records they complain that while they have satisfied themselves intellectually that their alternative thoughts are more realistic than their NATs, somehow the message doesn’t get through to them in a way that makes a real difference. This is the equivalent, in popular terminology, to the difference between head knowledge and heart knowledge.

However, if a suitable experiment can be devised to test out the claims of the negative thought or substantiate the truth of the alternative new position then often the impact is greatly reinforced. Indeed when in 2003 Bennett-Levy asked a group of CBT practitioners to compare the effectiveness of keeping thought records with using behavioural experiments to challenge their negative thinking the behavioural approach won hands down.

Why is experience our best teacher?

We do not fully understand why doing something rather than reading about it or learning it from someone else affects us so powerfully, but there are several theories that shed some light on the issue.

Firstly, there is a theory that when we have a direct experience of an event we lay down multiple ‘versions’ of the event in our brains. These might include motor or kinaesthetic representations associated with our body movements, the coding of visual and auditory information about the event and physiological feedback from our bodies – sensations, emotions and so on. The idea is that the combination of all these non-verbal memory traces lodges the meaning of the experience at a deeper level. We certainly know that events that stir our emotions are much more likely to be vividly remembered.

Secondly, memory and learning are also ‘context-dependent’. In other words, I stand a much better chance of recalling something if I can reproduce the environment and state in which I learned it originally. One experimenter discovered that a group of students who had committed a list of facts to memory when they were drunk did a better job of recalling them when they intoxicated themselves a second time than they did when they were sober! This means that lessons learned ‘in the field’, i.e. based on experiments in real-life situations that deliberately trigger your depressive or anxious thoughts, are much more likely to stick when you find yourself confronted by a similar situation in future.

Adult learning theories suggest that the most natural form of learning involves a continuous cycle of learning and reflection. This certainly fits with ideas about where our negative thoughts come from in the first place. We move through the following stages when making sense of new experiences:


When you think about it, human infants seem to be able to learn from and generalise from their experience long before they can be formally ‘taught’ or reason in any conscious fashion. If experience automatically stimulates us to ‘mentally digest’ its lessons and incorporate them into our internal working models of the world, then action and observation may be an indispensable part of deeper learning.

Why behavioural experiments are so helpful

You will have noticed by now that a great many of our negative thoughts rely on fear-filled assumptions – assumptions about what other people are thinking or indeed what will happen to us and others if we act in certain ways. Because we anticipate the worst we tend to play it safe and avoid situations and behaviours that might confirm our worst suspicions. But, of course, if our negative thinking is unrealistic and distorts reality, then there is a good chance that many of these gloomy or anxiety-inducing predictions are false. The great benefit of behavioural experiments is that they allow us to call the bluff of our negative thoughts and test out such predictions in real-life settings. These experiments can provide us with virtually irrefutable evidence that can bring even the most entrenched false belief crashing down.


Thomas used to suffer from panicky feelings every time he travelled on the underground. As a result he stopped using the tube and has to get around London by bus or bike. He was often late as a result. Subsequently he had developed an overwhelming fear that if he had a panic attack on the underground it would escalate until he went crazy. He had a mental image of himself rampaging down the carriage, screaming uncontrollably.

Eventually, having discussed the matter with his psychologist and learning why this was unlikely to happen, he agreed to put his belief to the test by taking a trip on the Circle Line in the company of a trusted friend who agreed to get him safely off the train at the next stop if there were any signs of him losing control.

What Thomas discovered is that he did feel highly anxious during the trip, especially during the initial stages. However, he neither went crazy nor lost control. In fact he found that his anxiety, although uncomfortable, was much more bearable than he had imagined. Having been reassured that his negative beliefs were untrue, Thomas was then gradually able to build up his tolerance of tube travel to the point where he was successfully able to use the underground without undue distress.

Behavioural experiments can also be used to trial behaviour based on freshly forged alternative beliefs and see whether or not such beliefs are indeed more helpful to us. What happens if I start acting in ways more in keeping with the goals I have set for myself, maybe doing some of the things I used to enjoy before I became so depressed? What are the consequences of striking up a conversation rather than protecting myself from potential rejection by hanging back and waiting for someone to come and talk to me? These are the sorts of questions you can put to the test using behavioural initiatives.

Dropping ‘safety’ behaviours that make things worse

According to CBT, because of the thought–feeling–behaviour links, changing what we do as well as what we think is vital. The way we act is often a powerful maintaining factor for our problems, not least because so many of the things we do to compensate for our problems end up making them so much worse. The compensatory strategies are referred to as ‘safety behaviours’.


To understand why these safety behaviours can be so unhelpful consider the case of Rose. Rose is a 35-year-old woman who is intensely shy in social situations. Among her underlying beliefs is the conviction that ‘people will think I am odd and I will be humiliated’. Usually she manages to avoid social gatherings altogether but when she can’t she has attempted to compensate for her belief that people will ‘see the fear in my eyes’ by wearing dark classes, even indoors. Rose is also very self-conscious about her tendency to blush when she feels insecure and so has taken to wearing very thick foundation to conceal any changes in her skin tone and a silk scarf that covers her neck at all times. Whenever she cannot escape conversations with new people, Rose relies on her trusty list of 10 stock phrases that she uses to ward off embarrassing silences.

Let’s consider the intended effects of Rose’s safety behaviours and their actual consequences in CBT terms:


What can be seen quite easily from this table is that a lot of Rose’s tactics actually end up having the very opposite effect to the ones she intended. This is very often the case with safety behaviours. Paradoxically, when we take our negative thoughts at face value and act on the basis of them we usually end up strengthening them.

It is really important to become aware of your own safety behaviours because you will need to learn to abandon them if they are keeping your problems going. Make sure your temporary fixes are not stopping you from sorting out your difficulties.

exercise4 icon

Take a moment to think about your problem and see whether you can compile a list of negative thoughts relating to it.

Looking at each thought in turn think about the ways in which it encourages you to behave. It might help to think about things that you would stop doing if your problem magically disappeared overnight.

Think about the potential problems that these behaviours might be making for you:

·        Do they stop you from confronting your fears?

·        How might they reinforce your negative beliefs?

·        Do they prevent you from exposing yourself to evidence that might contradict or undermine your NATs?

·        What effect do these behaviours have on other people?

·        How do these behaviours leave you feeling (a) in the short term and (b) in the long term?

·        In view of your understanding of the way in which thoughts, feelings and behaviours affect each other can you see any unwanted consequences of these behaviours?

My safety behaviours

Purpose or intention

Unwanted consequences





Devising behavioural experiments that work

Data gathering

There are two basic types of behavioural experiment. The first is where you put yourself in a situation to make observations and collect information. This might involve seeing whether you do actually react in the way you anticipate when you confront a scenario likely to bring your problem to the fore.

Often the most common safety behaviour of all, avoidance, means that we ‘protect’ ourselves from certain situations because we fear what will happen to us. Our negative thoughts tell us that exposure to the problem would be terrible for us and, of course, the more we stay away the bigger the problem becomes in our head.


Remember: every time you act in a way that is consistent with your negative thoughts you are unconsciously reinforcing the notion that they must be true!

Putting yourself into a real-life situation in which you believe you are likely to have difficulties is not easy, but does give you the opportunity to get a more realistic sense of your problem and the thought processes that underpin it. It also allows you to check out whether you really need to safeguard yourself against further exposure.

The other advantage is that you will probably have much better access to the relevant NATs and emotions, because the situation will trigger them, so you will then be able to target your intervention much more effectively. This can be invaluable for people who find it hard to recall what they thought and felt after the event, and it is interesting how often people come back with a completely different set of cognitions when you send them out to monitor their thoughts directly in a real-life setting.

Finally, you may also discover that the very process of putting yourself in the anxiety-provoking situation with the explicit purpose of monitoring your thoughts and feelings and logging their intensity may give you sufficient detachment to bring your stress levels down. Raising your self-awareness in this way keeps rational parts of your mind engaged and often stops the panic taking over.

Phone a friend?

The other form of data gathering relies on actively seeking out information and viewpoints that might help with the evaluation of negative beliefs. Someone with a phobia of flying may find it helpful to check out the actual statistics of the number of people killed in air accidents every year, or an obsessive person with irrational anxieties about catching AIDS may be reassured by researching hard facts about the ways the HIV virus is transmitted and how long it survives outside the body.

As you saw in the previous chapter, one of the most helpful counters for our biased negative thoughts is to think about the situation through the eyes of someone else, but sometimes it may be necessary to physically go and ask a sample of people for their views and perspectives. This can be particularly helpful with people prone to mind reading, who naturally tend to jump to all sorts of assumptions about what other people are thinking and feeling. Actually talking to them and finding out what they do think can expose just how far from the truth some of our more paranoid thoughts can be.

Prediction testing

The second type of experiment involves actively changing or manipulating something in order to test out the truth of a negative thought. This is the more traditional type of experiment where you check out a specific hypothesis, i.e. a theory about what will happen if your thought is true. If your experiment produces a result at odds with your prediction, you can be pretty sure that you need to take a long hard look at the original theory.

Turning your thoughts into predictions

Most of our negative thoughts carry implications that can be tested, but sometimes these need teasing out. If your negative thought is too vague or general it can be hard to put to the test. Experiments work best with predictions that are (a) quite specific and (b) will produce observable results. The formula to aim for is:

If X is true, then under these particular conditions Y should happen.

This does not mean that if your thought does not appear to be falsifiable through experimentation you should abandon this line of attack. What you may need to do is unpack the thought and test out some of its implications instead, thinking about what your belief would mean in practical terms across a variety of situations.

Say, for example, you held the negative belief: ‘I am a terrible person’. In its raw form this thought doesn’t lend itself easily to experimental investigation. But what would ‘being a terrible person’ look like in practical terms? What would it mean at work? At home? In the company of friends (assuming you are not such a terrible person that you don’t have any…)? What are the implications of being such a terrible person? What would a terrible person do – and what could they not do? On this basis you can start to formulate more focused predictions that you might be able to test out.


‘I can’t cope with life.’ This is what Vanessa, a 37-year-old administrator, had been telling herself over the months she had been sinking into depression. When she started to think about experiments that might help her question this belief, Vanessa recognised she needed to look at some of her related sub-thoughts and see whether there were any specific implications of her beliefs that could be tested. She came up with the following list:

Vanessa’s initial NAT: ‘I can’t cope with life any more

Implications/related thoughts



‘I have screwed up everything I care about’


Relies too much on subjective judgement

‘Past failures mean I will keep on failing’


Potentially falsifiable prediction

‘I have no reason to get out of bed any more’


Might warrant closer examination but not strictly testable

‘I have let everyone down’


Too much of a personal judgement but data-gathering survey might be useful to see if others agree

‘Even the smallest stress causes me to fall apart’


Potentially testable, providing ‘falling apart’ can be translated into operational terms

‘If I was at work I would make so many mistakes I would get sacked’


Could be tested but stakes are high and no opportunity to test while off sick. Might find some equivalent task and look at error rate?

Vanessa decided that the most promising of these was the implication that she would ‘inevitably fail in any new challenge’. This was a much more falsifiable proposition than her initial belief but nonetheless logically related to it in her own mind. The behavioural experiment she somewhat reluctantly decided upon was to undertake a photography course and test out her hypothesis regarding the inevitability of failure.

Despite her depression, and producing a portfolio of very bleak, austere black and white photographs for her final assignment, Vanessa turned out to have quite a talent for photography. Her efforts were highly acclaimed by her tutors and even she had to admit that, by any standards, she had indeed ‘succeeded’ in this new venture. Becoming more active and meeting new people on the course also helped alleviate some of her depressive symptoms and the net result was that her original negative belief ‘I just can’t cope with life’, which she had originally rated at 90%, dropped over 12 months to just 30%.


Decide how you are going to measure your results

When designing behavioural experiments do think carefully about how you are going to measure your results. Are there any objective benchmarks that will help you evaluate what has happened? If you rely on feelings or subjective impressions alone you may be in trouble: thinking errors such as filtering, exaggeration and polarised thinking could blunt your objectivity. Sometimes it can be helpful to recruit someone else to help you evaluate outcomes or at least develop firm criteria to which you can rate the outcome of your experiment.

Experimenting with safety behaviours

Once you have become aware of your safety behaviours, one of the most valuable experiments you can conduct is to see what happens if you drop them. The prospect of this is terrifying for most people because of the underlying belief that it is these behaviours that are protecting them from disaster.


A conundrum

A tribe performs an elaborate rain dance every year. And every year the heavens open and it rains. Members of the tribe naturally believe that it rains because their dance has appeased the spirits. Is there another explanation? More importantly, how would you be able to prove to the chief that the dance and the rain are not necessarily connected? Can you see a connection between the tribe’s attitude to the rain dance ritual and our relationship with our own safety behaviours?

Safety behaviours often allow us to create eminently testable predictions:

·        ‘If I don’t look directly ahead and keep reciting the Lord’s prayer I am bound to fall…’

·        ‘Unless I have a drink to relax me beforehand I will come across as awkward and tongue-tied…’

·        ‘If I don’t make myself think about something else whenever that picture comes into my head my bad thoughts will completely take over…’

·        ‘Unless I wear loose, baggy clothes every day people will think I am fat…’

·        ‘Unless I check my body in minute detail from head to toe every day I will certainly get cancer…’

With most safety behaviours the obvious experiment is usually to take your courage in both hands and see what happens if you abandon them. Will your negative predictions come true? If this feels too threatening you can also exaggerate safety behaviours and explore the consequences of doing so. If these behaviours are so helpful to you, won’t doing even more of them be even more useful? Or maybe not? Try it and see for yourself.

Challenging symptom-based predictions

These sorts of experiments can be very helpful for people who suffer from anxiety or any condition where one’s own physical and psychological reactions become part of a feedback loop that make things worse.

People who suffer from panic attacks often misinterpret their own anxiety symptoms as evidence that they are on the brink of catastrophe. For example, the release of adrenaline can produce a racing heart and sensations that can be misinterpreted as signs of an impending heart attack:

NAT: ‘My racing heart means I am going to have a heart attack…’

For such people deliberately accelerating their heart rate, say by running up a flight of stairs a number of times, can demonstrate that a rapid pulse can be a natural and harmless reaction to increased physical demands being placed on the body such as when it prepares itself for ‘fight or flight’. Unless they are predisposed to do so for some other health reason, the person concerned will not have a heart attack as the NAT predicts. In fact the acceleration should level off quite rapidly.

This knowledge will often help de-catastrophise the experience of a racing heartbeat the next time the person becomes anxious, and of course as they become more sanguine about the whole business a more relaxed attitude is less likely to prime a fight–flight response.

Similar experiments can be conducted with other symptoms such as dizziness. The belief that ‘feeling light-headed means I am about to lose consciousness’ can be readily disproved by hyperventilating, i.e. breathing really heavily for a couple of minutes. This should bring on quite a few of the sensations experienced during a panic attack but without the feared results associated with them.

Warning: If you do decide to experiment with physical symptoms, please make sure that you take appropriate medical advice. None of these procedures will harm a healthy person but it is worth talking to your GP and making sure you are in good health before you begin.

Experimenting with assumptions underlying your NATs

Sometimes you may find it hard to think of a way to test your NAT directly, but it is worth thinking about the background assumptions on which it relies, and seeing if you can devise a way to challenge those.


Kay had Obsessive Compulsive Disorder (OCD). She had become terrified that if she allowed herself to mentally picture her mother in a state of ill health then somehow magically she would bring about her death. She had developed all sorts of rituals to distract her and keep this image out of her mind.

We talked about the logical impossibility of there being any significant connection between thoughts in Kay’s head and her mother’s physical wellbeing. However, although she acknowledged that this made sense, the risk of inviting these images into her mind just seemed too high for her.

We decided to tackle her false belief indirectly by seeing whether Kay could use her supposed mental powers to affect the health of another living creature. The underlying assumption being tested was as follows:

‘What I do or don’t think can influence the wellbeing of other beings.’

Kay owned a small house with a garden of which she was extremely proud. The only problem was that the cat from next door would come and defecate in her flower beds. Over the next few weeks Kay decided that she would ‘will’ the unsuspecting cat into an early grave. Just using the power of her thoughts she would make this happen. Each day she would say out loud five times ‘Die cat, die!’ and she would picture the cat becoming sicker and sicker with each passing day. Unfortunately for Kay’s garden, the cat proved immune to her visualisations. Fortunately for Kay, the experiment was sufficient to introduce a chink of doubt into her conviction that her thoughts alone really could make her mother poorly.

What do I do if the results of my experiments actually end up supporting my negative belief?

The short answer is: don’t panic! There may be many reasons why things have turned out like this:

·        Check that your prediction really did target your negative thought.

·        Think about what other factors might be responsible for your results.

·        Did you design your experiment so that the result meant something? If you are testing the thought ‘nobody wants to know me’ and you get a rejection the first time you ask your friend round for coffee, that doesn’t really constitute an adequate test!

·        Don’t give up. Remember: you can never prove something is true (because the exception to the rule might still be lurking out there somewhere) but you only have to prove it false once. Keep challenging your negative thoughts and see whether you can come up with alternative ways of putting them to the test. It’s not easy, and it takes ingenuity and courage, but stick with it and the results will be worthwhile.

Homework: devise your own experiment

Looking at your thought records see whether you can turn your NATs and assumptions into potentially falsifiable predictions. If you are feeling brave, select one or two and see whether you can come up with an experiment to test them out.

If you are finding this difficult or too threatening, use your completed safety behaviours box on page 90 and see what happens if you deliberately abandon one of your usual strategies.

Remember: you need to work out beforehand how you are going to measure the outcome of your experiment.

If you find an error or have any questions, please email us at Thank you!