What a paper actually claims

Lesson 1 32 minReading

In this lesson

By the end, you’ll be able to

  • Separate results from claims from conclusions in a published abstract
  • Identify hedging language and match it to evidentiary strength

Sign in to save your progress

You can keep reading without an account, but completed lessons won’t be saved.

Sign in

A research paper is not a story told in the order the work happened. It is an argument, arranged so a reader can check it. Almost every empirical paper follows the same four-part shape, IMRaD: Introduction (why this question), Methods (what was done), Results (what was observed), and Discussion (what the authors think it means). The shape is a convention, not a law, but it is nearly universal in the sciences, and knowing it lets you find any part of the argument in seconds instead of reading front to back.

the fieldIntroductionnarrows to one questionMethods · Resultswhat was done, what happenedDiscussionwidens back out to the fieldwhat it means
IMRaD as an hourglass of scope: an Introduction that narrows from the broad field to one question, Methods and Results at the narrow waist, and a Discussion that widens back out to what the finding means.

Three things get confused constantly, and telling them apart is the whole skill of this lesson.

A result is what happened: the measured thing. "Participants in the treatment group scored 6 points higher on the anxiety scale."

A claim is what the authors say the result means about their study: "The intervention reduced anxiety."

A conclusion is what they say follows for the world beyond the study: "Brief interventions of this kind could be deployed in schools."

Each step travels further from the data. A paper can have a clean result, a defensible claim, and a conclusion the result cannot support. Most overreach lives in that last step, and most of it is invisible until you separate the three on purpose.

Worked example

Here is the move, worked once. Take the last line of the abstract you are about to dissect: “Our findings suggest that the timing of meals have far-reaching effects on hippocampal physiology and learned behaviour.” Nothing is measured in that sentence; it is a claim, the authors' reading of their results, and the hedge suggest tells you how far they will go. Now compare it to “mis-timed feeding resulted in dramatic deficits in hippocampal-dependent learning and memory”; that is a result, a specific thing they measured, in mice. Same paragraph, two different kinds of sentence. Read for the seam between them. (Source: Loh et al., eLife 2015, CC BY 4.0.)

Scientists mark how far they are willing to go with their language. "X causes Y" and "X may be associated with Y" are not stylistic variants; they are different sizes of claim. Words like may, suggests, is consistent with, and cannot rule out are hedges, and they are load-bearing. When you read an abstract, the verbs and qualifiers tell you how strong the authors themselves think the evidence is. A conclusion that drops the hedges its own results demanded is the single most common tell of overreach.

Exercise

A cohort study reports: "Adolescents in the highest tertile of nightly screen time slept 22 minutes less, on average, than those in the lowest tertile (95% CI 14–30 min)." The authors then write four sentences. Which one is a conclusion, a statement about the world beyond this study, rather than a result or a claim about the study itself?

Sort each sentence into result (what was measured), claim (what it means for this study), or conclusion (what follows for the world). Exactly one is a conclusion.

Answer options
How sure?

Signature interactive

The Claim Dissector

Span annotation

Here is a real abstract. Tag each part: what is background, what is a result the authors measured, and what is a claim they draw from it. Not every label is used, decide which are.

Background sets up why the question matters and what was done. Results are the specific things measured (clocks, LTP, CREB, memory deficits). A claim interprets those results, watch for the hedge "suggest". Conclusion and limitation do not appear here.

Pick a tag, then click the first word of a span.

How sure?

Source (CC BY 4.0): Loh DH, Jami SA, Flores RE, Truong D, Ghiani CA, O'Dell TJ, Colwell CS. Misaligned feeding impairs memories. eLife. 2015;4:e09460. CC BY 4.0. Retrieved via PubMed.

Written response

An observational study reports: "Teenagers who ate breakfast daily had, on average, GPAs 0.15 points higher than those who skipped it (95% CI 0.06–0.24)." In 3–5 sentences: state the strongest claim this result can support, and one conclusion it CANNOT support. Explain what stops the second one, using the right hedging.

Structure: (1) 'This result supports the claim that…' in associational, hedged language. (2) 'It cannot support the conclusion that… because…' and name one confounder or reverse-causation path.

0/40 words

How sure?

Lab

Below is a (constructed) press release about a real study. The paper's actual finding is shown first. Flag every sentence where the release claims more than the paper supports.

The paper. What the paper actually reports: in mice, feeding scheduled during the normal sleep period left the master clock (SCN) unchanged but shifted the hippocampal clock, reduced hippocampal long-term potentiation and CREB, and impaired hippocampal-dependent memory. The authors suggest meal timing has far-reaching effects on hippocampal physiology and learned behaviour.

Press release (constructed)

Real study: Loh DH, et al. Misaligned feeding impairs memories. eLife. 2015;4:e09460. CC BY 4.0. doi:10.7554/eLife.09460

Checkpoint · item 1 of 5

You want to repeat a study exactly. Which section tells you what the researchers actually did?

Which section is written so a stranger could run the study again?

How sure?

Reflection