Practice Drill: Conditional causal claims

Chinese version: /zh-hant/practice/drills/conditional-claims/

• Date: 2026-09-10
• Related goal: Current Goal
• Skill / subskill: Turn an overgeneralised causal claim into a conditional, testable statement.
• Why this drill: It tests whether the learner can move from “many factors exist” to a clear model that states variables, conditions, and a prediction.
• Starting level: fragile
• Timebox / repetitions: 5–10 minutes; one full response.

Performance condition

Without receiving a completed answer first, rewrite this claim in your own words:

「讀書時間越長，成績一定越好。」

Your rewrite must say what conditions matter. Then give one concrete situation in which merely increasing time would not be expected to improve the result.

Success criteria

• Does not use an unconditional word such as “一定” unless the conditions justify it.
• Names at least two relevant conditions or variables beyond time.
• States a directional relationship that could in principle be checked with evidence.
• Gives a counterexample or boundary case that is consistent with the rewritten claim.

Common error to watch

• Listing many factors without explaining how they change the original conclusion.

Attempt instructions

1. Write the response first, without asking for the final answer.
2. Save the unedited response under /practice/attempts/conditional-claims-attempt-01/.
3. Request feedback against the success criteria.
4. Write the smallest useful correction in a reflection note.

LLM coach prompt

Act as a deliberate-practice coach. Do not solve this task before I attempt it. After my attempt, identify the earliest reasoning failure, give one minimal hint, let me retry, then provide a rubric-based review. Update the recommended next drill based on my evidence.

Next-step rule

• If successful with a mechanism and counterexample: repeat with an unfamiliar domain.
• If partly successful: revise the claim by adding only the missing condition(s).
• If unsuccessful: practise distinguishing correlation, causation, and controlled conditions.