Running the Experiment and Observing Results
Do the planned activity, observe what actually happens, and record results immediately.
Key takeaways
- What it is: Do the planned activity, observe what actually happens, and record results immediately.
- Why it works: The emotional persuasiveness of a behavioral experiment comes from direct experience: not being told by a therapist that the belief is probably wrong, but experiencing the disconfirmation personally. This works through the same mechanism as exposure therapy — prediction violations at the level of experience are more belief-changing than at the level of reasoning. Recording results immediately matters because autobiographical memory is reconstructive and emotionally biased: memories formed hours after an anxious event tend to re-encode the event as more confirming of the anxiety than it actually was.
- Evidence: Backed by randomized trials / meta-analyses.
- Avoid: Running a diluted version of the experiment that doesn’t truly test the belief — wearing a safety behavior (e.g., speaking quickly in a presentation so people don’t notice nervousness) prevents the prediction from being genuinely tested.
Why it works
The emotional persuasiveness of a behavioral experiment comes from direct experience: not being told by a therapist that the belief is probably wrong, but experiencing the disconfirmation personally. This works through the same mechanism as exposure therapy — prediction violations at the level of experience are more belief-changing than at the level of reasoning. Recording results immediately matters because autobiographical memory is reconstructive and emotionally biased: memories formed hours after an anxious event tend to re-encode the event as more confirming of the anxiety than it actually was.
How to do it
- 1Do the planned activity as exactly as designed — do not run a safer version that doesn’t actually test the belief.
- 2During the experiment, notice the anxiety level (expected) and observe what is actually happening.
- 3Immediately after, write what happened — facts, not interpretations.
- 4Compare the actual results to the pre-stated criteria for confirmation and disconfirmation.
- 5Rate belief strength in the original prediction after the experiment (0–100%).
What the evidence says
RCT / meta-analysisBehavioral experiments are supported by exposure theory research: violations of threat predictions, when experienced directly rather than discussed, produce stronger and more durable belief change. A study by Salkovskis et al. found behavioral experiments produced greater belief change than verbal challenging alone for panic disorder.
Honest caveat: Evidence is primarily from anxiety disorders; generalization across all CBT applications of behavioral experiments is clinical rather than uniformly RCT-tested.
- — Salkovskis et al. (2007), "Belief disconfirmation versus habituation approaches in exposure therapy", Behaviour Research and Therapy
- — Bennett-Levy et al. (2004), "Oxford Guide to Behavioural Experiments in Cognitive Therapy"
Common mistake
Running a diluted version of the experiment that doesn’t truly test the belief — wearing a safety behavior (e.g., speaking quickly in a presentation so people don’t notice nervousness) prevents the prediction from being genuinely tested.
IX Coach tracks your pre- and post-experiment belief ratings across sessions, showing belief change over time and identifying which experiment types generate the most durable shifts for you.
Practice this with IX Coach →