Learn / Predictions you can check

Predictions you can check

3 minute read

The single most common reason an experiment fails to teach anything is a vague prediction. "It will go badly." "I will be awkward." "Something will happen." Afterwards, whatever occurred, the fear can claim it was right. A good prediction closes that door in advance.

Feelings are not outcomes

People with anxiety tend to be what the Oxford group calls emotional processors: they judge how a situation went by how it felt. It felt awful, so it went badly. But the feeling was there before anything happened, and it would have been there whatever happened. It cannot tell you about the world.

So the rule is: state the prediction as something a camera could record. Not "I will be boring," but "people will look at their phones, give one-word answers, and leave within five minutes." Not "I will panic," but "I will fall to the floor." Not "they will judge me," but "someone will comment on my face, or two people will exchange a look."

Questions that sharpen a prediction

How would we know if it happened? What would we see?
What would other people do or say?
How many people, how quickly, for how long?
What would count as it not happening?
What would disprove the fear?

Writing down the answer to "how would I know?" before you go is what lets you say, afterwards, "it did not happen," and believe yourself.

If, then

Predictions work best when they include the safety behavior you are going to drop. "If I speak without preparing, then people will think I am boring and stop talking to me." "If I keep standing when I feel faint rather than sitting down, then I will collapse." "If I send the email after one read-through, then it will contain a mistake my manager points out." The "if" names the change you are making. The "then" names the observable result.

The ratings

You rate two things, on a scale from 0 to 100.

The prediction. How likely does it feel that this will happen, this time? Rate it before, and again after. A drop from 70 to 20 is the experiment's headline result.

The belief. The broader belief the experiment is testing: "I am boring," "my heart cannot cope," "mistakes are unforgivable." Rate it when you first write it down, and re-rate it after each reviewed experiment. This is the number the app tracks over time.

The numbers are not precise measurements and do not need to be. They are a way of noticing change that would otherwise slip past. Beck and colleagues built ratings into cognitive therapy from the start for exactly this reason: conviction shifts in steps, and steps are easy to miss without a record.

Expectancy violation

Research on how fear is unlearned, led by Michelle Craske and colleagues, gives another reason to be precise. Learning is strongest when the gap between what you expected and what happened is large and unmistakable. That gap is called expectancy violation, and you can only notice it if you said clearly, beforehand, what you expected. The more specific the prediction, the sharper the surprise, and the more the brain updates.

The prediction is a bet you are making with your fear. Write it down so that afterwards there is no arguing about who won.

References

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.
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
Craske, M. G., Treanor, M., Conway, C. C., Zbozinek, T., & Vervliet, B. (2014). Maximizing exposure therapy: An inhibitory learning approach. Behaviour Research and Therapy, 58, 10–23. doi.org/10.1016/j.brat.2014.04.006
Hofmann, S. G. (2008). Cognitive processes during fear acquisition and extinction in animals and humans: Implications for exposure therapy of anxiety disorders. Clinical Psychology Review, 28(2), 199–210. doi.org/10.1016/j.cpr.2007.04.009
Beck, J. S. (2011). Cognitive behavior therapy: Basics and beyond (2nd ed.). Guilford Press.

Helpful links

Psychology Tools: How to use behavioral experiments

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