JH
Jonathan Haber

Run uncertainty experiments

Deliberately leave small things uncertain and observe what actually happens.

Key takeaways

  • What it is: Deliberately leave small things uncertain and observe what actually happens.
  • Why it works: Reassurance-seeking and checking are maintained by negative reinforcement: they produce short-term relief that prevents the disconfirmation of catastrophic predictions about uncertainty. Exposing yourself to real uncertainty — and not checking — tests whether the feared outcome (things go wrong, you cannot cope) actually occurs. Repeated disconfirmation erodes the belief that uncertainty is dangerous.
  • Evidence: Backed by randomized trials / meta-analyses.
  • Avoid: Selecting experiments that are too high-stakes initially — the brain’s threat system overwhelms the learning before the result can be registered. Start with genuinely trivial uncertainties.

Why it works

Reassurance-seeking and checking are maintained by negative reinforcement: they produce short-term relief that prevents the disconfirmation of catastrophic predictions about uncertainty. Exposing yourself to real uncertainty — and not checking — tests whether the feared outcome (things go wrong, you cannot cope) actually occurs. Repeated disconfirmation erodes the belief that uncertainty is dangerous.

How to do it

  1. 1Identify a low-stakes uncertain situation you would normally check or seek reassurance about.
  2. 2State your prediction: "If I don’t check, _____ will happen."
  3. 3Refrain from checking for an agreed period.
  4. 4Record what actually happened and compare to your prediction.

What the evidence says

RCT / meta-analysis

CBT protocols for GAD that include behavioral experiments targeting IU show significant reductions in worry and GAD symptoms in controlled trials.

Honest caveat: Effects are strong for GAD; generalization to normal-range worry is plausible but less directly studied.

References
  • — Dugas & Robichaud (2007), Cognitive-Behavioral Treatment for Generalized Anxiety Disorder — protocol evidence summary
  • — van der Heiden et al. (2012), RCT of uncertainty-based CBT for GAD, Journal of Consulting and Clinical Psychology

Common mistake

Selecting experiments that are too high-stakes initially — the brain’s threat system overwhelms the learning before the result can be registered. Start with genuinely trivial uncertainties.

IX Coach helps you design a graded set of uncertainty experiments at the right difficulty and tracks your prediction accuracy over time, surfacing the evidence that undermines IU beliefs.

Practice this with IX Coach →

Related practices