Taking action
How can I write one focused learning question for a small experiment?
Start with the next decision the experiment should inform, then choose one uncertainty that could change that decision. Write: “During [bounded action and context], what will I observe about [one uncertainty], and how will that evidence affect [the next decision]?” Make the answer observable, keep the scope small enough for the available time and resources, and state the decision use before you begin. If the question asks whether your whole direction is right, it is still too broad.
Begin with the decision, not the activity
Write the decision you expect to make after the experiment: stop, adapt, repeat, compare another option, or take one proportionate next step. Then ask what single uncertainty would most help with that decision. A task such as “try the class” is an activity; “Does the core practice fit the time and attention I can reliably give it?” is a learning question.
CDC guidance says evaluation questions define what will be investigated and should align with purpose, intended use, information needs, context, resources, and timing. That guidance is designed for program evaluation, not personal experiments. The useful transfer is narrower: a learning question should earn its place by helping with a stated decision.
Use the FOCUS question card
FOCUS is a Forever Free Compass framework, not an external research finding. Fill the five fields before you begin the small experiment.
- Future decision: name the literal choice the answer should inform.
- One uncertainty: select one assumption or unknown that could change that choice.
- Conditions: bound the action, setting, people involved, time window, cost, and other relevant limits.
- Usable observation: state what you can notice, record, or ask without pretending that it proves more than it does.
- Shift rule: write how a clear yes, no, mixed, or unusable answer would change the next step.
Turn a broad hope into an answerable question
Use this sentence frame: “During [bounded action and context], what will I observe about [one uncertainty], and how will that evidence affect [next decision]?” You may shorten it, but keep every function visible somewhere in the plan.
The 2026 Test and Learn guidance recommends identifying critical assumptions, agreeing decision criteria before testing, and selecting methods proportionately to the question, decision, time, and resources. It also separates exploratory testing from later evaluation. For a personal experiment, that means asking for useful learning rather than proof that an entire direction will succeed.
| Broad version | Focused version | Decision use |
|---|---|---|
| Is this direction right for me? | During two 45-minute practice sessions this week, what recurring friction and usable interest do I record about the core activity? | Decide whether to stop, redesign the conditions, or run one more bounded test |
| Will this idea work? | When I show one rough example to three intended users, which part do they understand without explanation and where do they hesitate? | Choose what to revise before building more |
| Can I keep doing this? | Across four scheduled starts, what prevents or enables me to begin within ten minutes? | Keep the cue, change the setup, or choose a smaller action |
Match the question to the stage of learning
An early experiment can ask about feasibility, delivery, comprehension, participation, or one proposed mechanism. It should not quietly become a claim that the whole approach caused a lasting outcome. CDC distinguishes formative questions used during development from process, outcome, and impact evaluation questions used for different purposes and stages.
The Test and Learn guidance likewise says early testing should use methods appropriate to the question and should not turn immature evidence into a strong conclusion. If your question contains words such as “prove,” “always,” “guarantee,” or “the best,” rewrite it around what this bounded action can actually reveal.
Pre-commit to the observation and its use
The Cabinet Office playbook describes a test as an assumption and related hypothesis, an activity, and a way to observe and record results. It also recommends using evidence to update assumptions and guide the next decision. On a FOCUS card, the Usable observation and Shift rule do that work in plain language.
Write the observation before the action so a disappointing result is not redefined after the fact. Include an “unusable” branch: if the action was not completed, the conditions changed materially, or the observation could not answer the question, record that limitation and redesign the test instead of forcing a conclusion.
Keep the experiment within its authority
A focused question does not make an experiment safe, ethical, or decisive. Do not use a personal test to bypass consent, confidentiality, professional duties, required processes, or significant health, legal, financial, employment, housing, or safety consequences. Use qualified guidance and the relevant formal process when the stakes require it.
The honest outcome may be that the question was too broad, the observation was too weak, or the answer remains mixed. That is still learning. Narrow the uncertainty, change the conditions, gather a different kind of evidence, or decide that another experiment would not be proportionate.
Original contribution
The Forever Free Compass take
Forever Free Compass uses the FOCUS question card as an original reflection framework: Future decision states what the answer will inform; One uncertainty selects a single assumption; Conditions bound the action, context, and observation window; Usable observation names what can actually be noticed or recorded; Shift rule states how different answers would change the next step. FOCUS is a low-stakes question-writing aid, not a validated research, evaluation, experimental-design, causal-inference, or decision method, and it cannot make weak evidence conclusive.
Sources
- Step 3 – Focus the Evaluation Questions and Design — Centers for Disease Control and Prevention. Question boundaries; purpose and intended use; information needs; context; resources; timing; formative, process, outcome, and impact question types.
- Test and Learn — UK Government Evaluation Task Force. Critical-assumption testing; advance decision criteria; proportionate methods; early learning; separation of exploratory testing from robust later evaluation.
- Testing, learning and iterating (phase 4) — UK Cabinet Office. A test structure comprising an assumption and hypothesis, an activity, and an observation record; using evidence to guide continuation, iteration, pausing, or stopping.
This is educational reflection, not medical or mental-health care. If distress is persistent, severe, or involves immediate safety, contact a qualified professional or local emergency service.