Choose the response
5 min · Turn disagreement into evidence
Preserve the initial assumptions and record what evidence should trigger reconsideration. Re-estimation should expose learning rather than erase history.
Reading
coreMatch response to cause
Rewrite a misunderstood boundary, expose a hidden dependency, compare competing approaches or run a bounded investigation for missing evidence. Then re-estimate, split the work or retain a range. Record the original assumption and a reconsideration trigger.
This four-part response is an authored synthesis, not a named industry standard.
Worked example and transfer practice
All examples below are fictional teaching material. Use synthetic notes; no workplace or personal data is required or uploaded.
End an investigation with a decision. The fictional supplier connector is unknown. The team timeboxes a test using synthetic records, checks authentication and rate limits, and records whether the minimum export can run. The exit is an evidence review, not ‘keep investigating until comfortable’. The result may support re-estimating, splitting the work, retaining a range or dropping the option.
Try before reading the guide: The test succeeds for a small batch but fails at expected volume. State the limitation, one next option and a condition for reconsidering it. Keep both the original assumption and the new evidence.
Practice review guide
Do not treat the small-batch result as proof of production scale. A reduced batch option or another bounded test can be defensible if its outcome, risk and owner are explicit.
Knowledge check
corePractice · formative check
Choose the response
Select the response that preserves uncertainty honestly.
Concepts: uncertainty, learning action
There is no attempt limit or pass/fail state. Completing the check only enables the lesson's separate explicit completion control; it never records completion or certificate eligibility itself.
Activity ready. Attempts are unlimited. Completing this practice activity enables the lesson's separate, explicit completion control.
Equivalent non-interactive path
Answer and feedback guide
The team cannot size an integration because the vendor API behaviour is unknown. What is the best next action?
- Add a large private buffer and report one number: Hidden padding preserves neither the uncertainty nor its trigger.
- Run a bounded investigation and define what evidence will end it (best answer): Correct. Resolve the material unknown before manufacturing precision.
- Size only the known work and omit the dependency: Omitting a material dependency makes the estimate misleading.
Sign in to save progress across devices.Sign in
Course outlineLesson 3 of 3
Turn disagreement into evidence