Learn / Behavioral experiments

Behavioral experiments

6 minute read

A behavioral experiment is a planned activity, carried out in real life, designed to test a belief. You say what you expect to happen, you do the thing, and you compare what you predicted with what you observed. The Oxford Guide to Behavioural Experiments, which set out the method in detail, calls them among the most powerful tools in cognitive therapy, because they let people learn through direct experience rather than argument.

Experiments are not exposure

This distinction matters. Traditional exposure therapy asks you to stay in a feared situation until the anxiety fades. Repetition does the work. A behavioral experiment asks a different question: is the thing I am afraid of actually true? The goal is new learning, not a lower anxiety score. Anxiety may or may not fall during the experiment, and that is beside the point. What matters is whether the prediction came true.

One practical consequence: you rate the prediction, not the anxiety. Repeatedly checking how anxious you feel keeps your attention on your body and your feelings, which are exactly the unreliable witnesses you are trying to get away from. The exception is when the belief is about anxiety itself, such as "if my anxiety gets too high I will lose control." Then the anxiety is the thing under test.

A systematic review comparing behavioral experiments with exposure alone found some evidence that experiments do better, though the number of direct comparisons is small. What the review is clear about is that the two are different procedures with different aims, and they should not be confused.

Two kinds of experiment

Discovery experiments show you how a process works, without a fixed prediction. Try not to think about a green rabbit for one minute and notice what happens. Hold a conversation while monitoring yourself closely, then hold another while focusing on the other person, and compare. Read pairs of words like "dizzy, faint" and notice what your body does. These teach you, from the inside, that thoughts produce sensations, that suppression backfires, and that where you put your attention changes what you notice.

Prediction-testing experiments are the main kind, and the kind the app is built around. You name a specific, observable prediction, rate how much you believe it, arrange a situation that tests it, drop the safety behaviors that would let you explain the result away, and then see.

A third variety deliberately produces the feared outcome to test its cost. Say something boring on purpose. Wear a little blush to the shade you feel your face goes. Let your hand shake while holding a cup. Then watch how people actually react. These are usually done with a therapist first.

The seven steps

Warnock-Parkes and colleagues, drawing on decades of work at Oxford, describe seven steps.

1. Identify the belief. Choose one that is tied to strong emotion, gets in the way of life, and matters to you. If several qualify, start with the one most likely to give an early success. 2. Find the best situation to test it. Not the scariest. The best test you are willing to try. Real situations, out in the world, beat imagined ones, and many take only a few minutes. 3. Make the prediction observable. "I will be boring" is not testable. "People will look away and stop talking to me within two minutes" is. 4. Name the safety behaviors and other maintaining processes to drop. Including where your attention will go. Hands off the wall. 5. Do it, and watch. Attention outward. Notice what actually happens, not how it feels. 6. Review the outcome and draw the general lesson. What happened? Did the prediction come true? What does that say about the belief more broadly, about how you come across, about how safe you actually are? 7. Build on it. Listen for the "yes, but…" and turn it into the next experiment. Repeat, scale up, vary.

When the feared thing happens

Sometimes it does. Someone does laugh. You do stumble. The experiment has not failed. You now get to test the second half of the fear: was it as bad as you expected? How did people actually respond, and for how long did it matter? People usually find the real cost is far smaller than the imagined one, and that is often the more important discovery.

References

• Bennett-Levy, J., Butler, G., Fennell, M., Hackmann, A., Mueller, M., & Westbrook, D. (Eds.). (2004). Oxford guide to behavioural experiments in cognitive therapy. Oxford University Press. doi.org/10.1093/med:psych/9780198529163.001.0001
• Rouf, K., Fennell, M., Westbrook, D., Cooper, M., & Bennett-Levy, J. (2004). Devising effective behavioural experiments. In J. Bennett-Levy et al. (Eds.), Oxford guide to behavioural experiments in cognitive therapy (pp. 21–58). Oxford University Press.
• Bennett-Levy, J. (2003). Mechanisms of change in cognitive therapy: The case of automatic thought records and behavioural experiments. Behavioural and Cognitive Psychotherapy, 31(3), 261–277. doi.org/10.1017/S1352465803003035
• McMillan, D., & Lee, R. (2010). A systematic review of behavioral experiments vs. exposure alone in the treatment of anxiety disorders: A case of exposure while wearing the emperor's new clothes? Clinical Psychology Review, 30(5), 467–478. doi.org/10.1016/j.cpr.2010.01.003
• Warnock-Parkes, E., Thew, G. R., Murray, H., Grey, N., Wild, J., Kerr, A., Smith, A., Stott, R., Ehlers, A., & Clark, D. M. (2025). Driving cognitive change: A guide to behavioural experiments in cognitive therapy for anxiety disorders and PTSD. Cognitive Behaviour Therapy, 54(6), 767–793. doi.org/10.1080/16506073.2025.2518427
• Salkovskis, P. M., Clark, D. M., & Hackmann, A. (1991). Treatment of panic attacks using cognitive therapy without exposure or breathing retraining. Behaviour Research and Therapy, 29(2), 161–166. doi.org/10.1016/0005-7967(91)90044-490044-4)

Helpful links

• Psychology Tools: How to use behavioral experiments
• Centre for Clinical Interventions: Stepping out of Social Anxiety (Modules 3 and 4)
• Warnock-Parkes et al. (2025), open access

Educational content, not medical advice or a substitute for professional care. Ready to try it? Open the online worksheet or get the app.