CBT guide · Test a prediction

Behavioral Experiment

A behavioral experiment turns a prediction into something you can observe. Instead of arguing yourself out of a thought, you try a small action, see what happens, and revise the prediction only as far as the result warrants.

← Back to The CBT Pattern Map

Guided practice

Walk through an example

Select each step to see how it works, then try the practice question.

Make the prediction specific

Name the action, what you expect to happen, and how likely it seems. Translate a claim about what someone thinks into a response you could actually observe.

The prediction

Casey predicts an 80% chance that asking a coworker to clarify an instruction will lead to a critical remark or a refusal to help.

Plan a small test

Pick a situation where you can take the action and agree in advance what would count for or against the prediction.

The test

During a routine handoff, Casey will ask one clear question about which version of a file to use and listen to the response.

Observe what happens

Notice what was said or done. Keep the observation separate from your interpretation and from how nervous you felt.

The result

The coworker points Casey to the right file and moves on. Casey still feels nervous, but hears no criticism and gets an answer.

Update the prediction

Compare the outcome with the prediction. If the results were mixed or unclear, say so. Decide what to change about the prediction or what to test next.

What changed

“This coworker answered this question without criticizing me. The result doesn’t tell me what everyone will do, but my prediction wasn’t right this time.”

Which plan gives Casey the clearest test?

  1. Ask one routine clarifying question and notice how the coworker responds. This creates an opportunity for the predicted response and gives Casey something specific to observe. Feeling nervous doesn’t count as evidence that criticism occurred.
  2. Look up the answer alone and see whether the worry goes away. Looking it up might solve the file question, but it doesn’t test what happens when Casey asks. The prediction remains untested even if the worry settles.
Think through your own situation

You don’t need to enter anything here. Use the prompts on your own.

  1. What do you expect to happen if you take one specific action?
  2. What small test would give that outcome a chance to happen?
  3. What could you see or hear that would count for or against the prediction?
  4. What did the result show? What questions did it leave open?

When to use it

Use it when the thought takes the form “If I do this, then that will happen,” and you can test it through a small, ordinary action.

A test can be useful even when the feared outcome happens. Notice what actually happened, how you handled it, and which parts of your prediction still need testing.