Plan a behavioral experiment

Plan a small safe test of a prediction with a clear observation and review plan.

How to use it

Work with a manageable everyday situation. The sequence below is an invitation, not an assessment. You can skip by ending the exercise, correct your wording, or stop at any time.

1. What specific everyday prediction would you like to test?
2. What low-risk action could give useful information? Do not test physical danger, medical advice, or another person's consent.
3. What would you observe? What would count for or against the prediction?
4. When could you try it? What support, stop condition, and review time would help?

With mira

Open this technique in mira. Review the selection before starting. This scripted exercise can run offline; it makes no AI requests. Your answers are saved on this device in the session's encrypted exercise state. They are not automatically inserted into ordinary AI chat.

Scope and source

An original self-reflection adaptation of CBT principles, not a complete treatment protocol or a clinically validated chatbot intervention. It cannot assess risk. Avoid overwhelming or traumatic material; stop if distressed and consider qualified support. Acceptance never means accepting abuse, danger, or an unwanted situation.

Judith S. Beck, Cognitive Behavior Therapy: Basics and Beyond, 3rd edition (2020). Beck Institute resources.

Executable Pattern Markdown

The same version-1 source is bundled in mira. This view is for inspection; it does not execute automatically.

:language 1
:title "Plan a behavioral experiment"
:output outcome
= focus "Plan a small safe test of a prediction with a clear observation and review plan."
!use therapy.ready as readiness {
    :with focus
}
?branch {
    :when readiness.decision.value == "begin" -> explore
    :when readiness.decision.value == "support" -> support
    :else -> declined
}
explore:
!ask step1 {
    What specific everyday prediction would you like to test?
}
!ask step2 {
    What low-risk action could give useful information? Do not test physical danger, medical advice, or another person's consent.
}
!ask step3 {
    What would you observe? What would count for or against the prediction?
}
!ask step4 {
    When could you try it? What support, stop condition, and review time would help?
}
!show {
    Your reflection, in your own words:

    1. {{step1}}
    2. {{step2}}
    3. {{step3}}
    4. {{step4}}

    You may leave this here. Nothing needs to feel resolved.
}
= outcome "reflected"
-> end
support:
!show {
    Stop the exercise. If anyone is in immediate danger, contact local emergency services. Reach a trusted person or qualified crisis support if you can. This exercise cannot provide emergency care.
}
= outcome "support_requested"
-> end
declined:
!show {
    We can leave this here. You do not need to explain or finish.
}
= outcome "declined"
-> end