Reading a Primary Paper: The Claim, the Figure and the Missing Control
Trace a paper's central claim to its measurements and comparisons. Check controls, sampling and analysis choices, then distinguish observed results from interpretations and wider conclusions.
What a learner can do afterwards
- State a paper's central claim in one sentence and identify the figures, tables or analyses that test it
- Identify a control that is absent and say which alternative explanation it would have excluded
- Separate the results the data support from the conclusions the discussion asserts
1 · Read
A primary paper reports a study. The abstract summarizes its question and main findings. Methods describe design and analysis; results present measurements and analyses; discussion interprets them. These sections can overlap. A heading does not establish that an assertion is correct. Use methods and legends to check how reported outcome values were obtained.
State the central claim, then locate the figures, tables or analyses that test it. There may be several. Identify the comparison, outcome, population and follow-up. The largest sample or smallest p-value need not address the central claim. Use supplementary material when necessary details are absent from the main figure.
Symptoms fall after a molecule is given in a vehicle solution. A matched vehicle-only group would help distinguish the molecule from vehicle, handling and change over time. Random assignment and reliable outcome measurement still matter. Even a randomized comparison showing lower symptoms at day 7 does not establish a lasting cure without suitable follow-up.
Separate a sample observation from a wider inference. Lower scores in the measured sample do not establish benefit for all patients. Results can contain statistical inferences, and discussion can cite evidence. Judge each assertion by its support, assumptions and uncertainty, wherever it appears.
Apply the previous topics to legends and methods. Check what n counts, whether observations share donors or animals, and how grouping and error bars were handled. Compare reported outcomes, exclusions and stopping with the registered plan when available. An outcome selected after inspecting results should be identified, with the deviation and exploratory status reported.
For a proposed mechanism, name the additional measurement or comparison needed to test it. Missing evidence alone does not establish misconduct. When papers appear to disagree, compare outcomes, populations, designs and uncertainty before treating them as tests of the same claim.
Trace claims to relevant comparisons. Check controls and analysis, then judge each conclusion's scope, assumptions and uncertainty wherever it appears.
Original lesson, questions and flashcards about evaluating research claims, written for this learning dataset. Study scenarios are hypothetical.
The four readings and accompanying flashcards were written for this learning dataset.
More to read
Map the claim to the comparison
Write the central claim in one sentence. Identify its outcome, population, comparison and follow-up. Then find the figures, tables or analyses that address those features. Several pieces may be needed. A large sample studying a different outcome does not answer the claim. Use methods, legends and supplementary material to recover missing design details.
Make a control criticism specific
Name the alternative explanation before proposing a control. For an injected molecule, vehicle and handling accompany the molecule. A matched vehicle-only injection group helps separate those conditions from the molecule's contribution. Assignment and outcome measurement still need scrutiny. A useful control addresses a named alternative; it does not make every broader conclusion valid.
Check how far the conclusion extends
A sample observation differs from a prediction about other people, settings or times. Ask what was measured and where the conclusion extends beyond it. Results may contain statistical inferences, while discussion may cite relevant evidence. Section headings do not settle credibility. Evaluate the support, uncertainty and assumptions behind each assertion.
Read design and reporting together
Check what n counts and whether points share an animal, donor or batch. Read how the analysis handles that grouping and defines error bars. If a registered plan exists, compare its outcome, exclusions and stopping procedure with the report. Ask for the timing and reasons behind changes. An unexplained difference requires investigation, not an automatic accusation of misconduct.
2 · Watch
Take it off screen
Where it sits
Where this leads
8 questions wait behind this lesson, each with its answer explained. Every answer feeds the sky: stars light as they are learned, and dim when it is time to come back.