Read
The Claim Is Not the Source
1. A convincing folder
At 14.10, the editorial team thinks the difficult work is over. Their article concerns a proposed change to the school day. On the shared screen sits a folder: a policy document, a student survey, an interview transcript, a photograph of a crowded corridor, and two opinion posts. The files have titles, dates, and working links. Nobody has fabricated a document. Nevertheless, the team still does not know whether its main sentence is defensible: “The proposed timetable will improve learning for most students.” A folder full of genuine material can create the feeling of certainty before the evidence has earned it.
2. Begin with the exact claim
The editor, Noor, asks everyone to stop discussing whether the sources are “good”. Good for what? A source has no universal usefulness. The team rewrites its broad sentence as three smaller claims: the proposal changes when lessons begin; later starts may alter how rested some students feel; and the change will improve learning for most students. The policy document can establish the first claim if it is current and official. It cannot, by itself, establish the second or third. By making each claim precise, the team can ask what kind of evidence would actually be needed.
3. Provenance answers one set of questions
Provenance is the traceable history of a source: who created it, when, where it appeared, for what purpose, and whether it has been altered or removed from context. The team’s policy document has clear provenance. It comes from the body responsible for the proposal and states what administrators intend to do. This makes it strong evidence for the content of the proposal. It does not make the document neutral, complete, or capable of predicting the results. Authority is connected to a domain. The author of a timetable can describe the timetable without automatically becoming an expert on every effect it might have.
4. Usefulness depends on claim-evidence fit
Claim-evidence fit describes how directly a piece of evidence supports a particular statement. A source may be authentic, informed, and still be a poor fit. The student survey asks whether respondents “prefer a later start”, but the article’s claim concerns learning. Preference and learning are not identical. A student might prefer the change because travel becomes easier, yet learn no more. Another might dislike the change but concentrate better. The survey may support a carefully limited claim about the surveyed students’ stated preferences. Using it to prove improved learning would stretch the evidence beyond what was measured.
5. Read methods, not only results
The survey contains another warning. It was shared through one student group’s social channel and completed voluntarily. The team cannot assume that the respondents represent the whole school. Perhaps students with strong views were more likely to answer. Perhaps some year groups barely saw the link. These possibilities do not make the survey worthless. They change the size and wording of the claim it can support. “Among the students who answered our voluntary survey, a majority preferred the later start” is more defensible than “Students prefer the later start.” Careful qualification is not verbal decoration. It records the boundary of the evidence.
6. Expertise is specific, not magical
The interview is with a teacher who has worked at the school for many years. The teacher offers detailed observations about tired classes and rushed mornings. Experience can reveal patterns worth investigating, but it does not automatically show what caused them. The teacher sees students in a particular subject, at particular times, and through a professional role. The interview is useful for describing a perspective and identifying possible mechanisms. It would be weak evidence for a numerical claim about all students. Calling someone an expert should open two questions, not close the discussion: expert in what, and speaking from what evidence?
7. Corroboration must add independence
The team finds three posts repeating the same statement about later starts. At first, this looks like corroboration, the confirmation of a claim through additional evidence. A closer check shows that all three posts copied the statement from one anonymous message. Repetition has increased, but independence has not. Ten versions of one unsupported claim do not become ten sources. Stronger corroboration comes from evidence with a separate route to the fact: an original document, a person with direct knowledge, a dataset collected independently, or another investigation that explains its method.
8. Disagreement can be useful evidence
Two sources may also conflict without one being dishonest. The administrator says the change has been designed to reduce disruption. A student representative argues that late activities will create new problems. They may be evaluating different outcomes or speaking about different groups. Instead of forcing an immediate winner, the team maps the disagreement: What exactly does each person claim? What access do they have? Which interests shape their attention? What evidence could distinguish the claims? A contradiction can expose hidden definitions, missing groups, or assumptions that a smooth consensus would leave invisible.
9. Images make claims too
The corridor photograph feels powerful because the crowd is visible. Yet an image has a frame. The photograph shows one place during one short moment. It does not show whether the corridor is usually crowded, what happened immediately before, or whether the timetable proposal caused the scene. The image can document that the corridor was crowded when the photograph was taken, if its date and context are known. A caption that calls it “daily chaos caused by the current timetable” would add claims the pixels cannot establish. Visual evidence must be sourced and interpreted with the same care as prose.
10. Lateral reading changes the direction of attention
Inspecting one page's design, language, and credentials can help, but polish does not guarantee strong evidence. Lateral reading means leaving the source temporarily to investigate it through other sources. The team checks who runs a campaign site, whether the policy document appears on the responsible body’s own pages, and whether a dramatic quotation comes from a longer statement. The aim is to understand the source’s position in a wider information system.
11. Uncertainty has different causes
By late afternoon, the team knows more, but it also sees more clearly what it does not know. Some uncertainty comes from missing evidence. Some comes from limited methods. Some comes from a future outcome that cannot yet be observed. These are different problems. Missing context may be recoverable. A narrow sample may require a narrower claim. A prediction may need scenarios rather than certainty. The phrase “we cannot know anything” would be as misleading as the original confident headline. Responsible uncertainty identifies what is supported, what remains open, and what would change the judgement.
12. Publication is a chain of decisions
The team replaces its main sentence: “The proposal would move the start of the school day, and many respondents to our voluntary survey welcome that change; its effects on learning remain uncertain.” The new sentence is less dramatic, but more informative. It separates a documented feature, a bounded finding, and an unresolved prediction. The article can still explain why supporters expect benefits and why critics anticipate costs. It can show readers where the evidence is direct, where it is testimonial, and where it is provisional. Accuracy is not achieved by removing every interpretation. It comes from making the route from evidence to claim visible enough to examine.
13. Pressure does not excuse distortion
Editors often work with incomplete information, limited time, and audiences that reward simple conclusions. Those pressures do not change what a source supports. The response may be to reduce the claim, mark the limit, postpone publication, or explain what remains unverified. Source evaluation is not a checklist added after writing. It shapes language: which verb to use, how broad the subject can be, whether a number needs context, and where uncertainty must be visible.
14. The source is only the beginning
A credible name, a working link, or an official logo may help establish provenance. None of them performs the reasoning for you. Before using evidence, state the claim precisely. Ask what the source can know, inspect how its information was produced, seek genuinely independent corroboration, and compare the strength of the evidence with the strength of your wording. The central habit is simple to state and difficult to practise: do not ask only, “Can I trust this source?” Ask, “What, exactly, can I responsibly claim because of it?”
> Keep this idea: Source evaluation is a relationship between a claim, a piece of evidence, a method, and a purpose.
