Reading a Primary Paper: The Claim, the Figure and the Missing Control · seed 1 · A4, ink-friendly. The answer key prints on its own page for grown-ups.
En anglais

Reading a paper for what it really shows

Science · Scientific Inquiry · ages 22-24
Name ______________________   Date ____________

  1. Which step connects a central claim to its evidence?

    • Choose the largest sample even when its outcome differs from the claim
    • Locate comparisons that address the claim
    • Choose the smallest reported p-value even when it concerns another population

  2. You want the measured outcome values behind a paper's claim. Where would you usually start?

    • Results figures and tables
    • The abstract's summary conclusion without its supporting values
    • The methods' recruitment target without the observed outcomes

  3. Every assertion in a discussion section was directly measured in the study.

    Circle one:   True   False

  4. A report says 'treated participants had lower scores' and 'the treatment will help all patients'. What distinguishes these statements?

    • A sample result versus a wider inference
    • Two sample descriptions because both mention the same treatment
    • Two direct measurements because both appear beside the results figure

  5. Symptoms improve after a molecule is injected in a vehicle solution. Which comparison best addresses vehicle and injection-handling explanations?

    • Untreated animals receiving neither vehicle nor the injection procedure
    • Animals receiving the molecule in a different vehicle with different handling
    • A matched vehicle-only injection group

  6. A registered whole-sample outcome is omitted. After inspecting results, authors select a significant subgroup outcome and call it primary. What should a reader request?

    • Only the subgroup p-value, because significance establishes that it was the intended primary outcome
    • The planned result and a dated explanation of the change
    • Only a larger subgroup sample, because more observations would resolve the reporting discrepancy

  7. A study measures symptom scores but proposes an unmeasured molecular mechanism. How should that proposal be assessed?

    • As an inference needing a mechanism test
    • As directly measured because its explanation fits the symptom pattern
    • As established because the symptom comparison reached statistical significance

  8. A randomized fertilizer study measures taller seedlings after two weeks. Its abstract claims greater harvest yield. Which limitation remains?

    • Randomization turns early height into a direct measurement of harvest yield
    • The control comparison removes the need to observe outcomes later in the season
    • Early height does not establish harvest yield
LightMySky · lightmysky.comW1-mt_DZE7z5vu9X-s1

Answer key

For grown-ups. Fold this page away before handing over the rest.

Reading a paper for what it really shows W1-mt_DZE7z5vu9X-s1

  1. Locate comparisons that address the claim · Relevant comparisons matter more than figure size or statistical extremity alone.
  2. Results figures and tables · Results usually present the measured outcomes; methods and legends help check how they were obtained.
  3. False · Discussion can interpret measurements or extend beyond them. Check each assertion's support.
  4. A sample result versus a wider inference · The first concerns measured participants; the second extends to people outside that sample.
  5. A matched vehicle-only injection group · The matched group retains vehicle and injection handling while omitting the molecule.
  6. The planned result and a dated explanation of the change · Readers need the planned result, the deviation and its timing, with the selected analysis identified as exploratory.
  7. As an inference needing a mechanism test · Symptom measurements alone do not directly test the proposed molecular process.
  8. Early height does not establish harvest yield · The measured outcome and follow-up differ from the harvest-yield claim.
Worksheet · LightMySky