Evaluating Claims, Evidence & Arguments

Map claim, reasons and counterclaims

Start by stating exactly what the writer wants accepted. A proposal (“extend library hours”) is not identical to an empirical prediction (“the extension will increase access for shift workers”). Then label the offered evidence and the reasoning that connects it to the claim. An unstated assumption often sits in that connection.

Original example: “The evening survey received 80 requests for later hours, so the library should close two hours later every day.” The survey supports interest among its respondents. Moving to a daily policy additionally assumes that demand is representative, sustained and worth the cost. A counterclaim offers a competing position; a rebuttal responds with reasons or evidence rather than merely repeating the proposal.

Distinguish disagreement from evidence of error. A person may accept that later hours help some residents but oppose the cost. That is a value or priority difference, not necessarily a dispute about the survey count.

Test relevance, sufficiency and source quality

Relevance: does the evidence address this exact claim? Awards for a building's design do not establish demand for evening access. Sufficiency: is there enough representative evidence? Two vivid anecdotes can illustrate an issue without estimating its prevalence. Quality: how were observations collected, when, by whom, and with what incentives or limitations?

A primary source is close to an event or data collection, but it can be mistaken or incomplete. A careful secondary synthesis may offer stronger support for a broad conclusion. A source's expertise matters only where relevant; a title or institutional logo does not establish every claim. Corroboration is stronger when sources are genuinely independent, not five websites repeating one press release.

Worked example: a percentage needs a denominator

A club says attendance increased 100%. Its log shows an increase from 2 to 4. The arithmetic is correct, but the absolute gain is two people. This alone is weak support for renting a hall for 200. Check the baseline, sample size, comparison group and time window before treating a percentage as decisive.

Evaluate reasoning and strengthen fairly

Common weaknesses include confusing sequence with cause, generalizing from a selected sample, attacking a speaker instead of the evidence, or presenting only two options when alternatives exist. These labels are useful only when tied to the actual reasoning. An emotional example is not automatically false; it may illustrate an experience while failing to establish how common it is.

For “strengthens” questions, choose evidence that repairs the argument's important gap. If the gap is representativeness, a broader sample helps more than a famous endorsement. For “weakens” questions, identify information that undermines a necessary connection, not an unrelated negative fact. A counterexample can refute an “all” claim while leaving a narrower “some” claim intact.

Balanced evaluation

Claim: opening a second entrance reduced waiting time. Before-and-after waits fell from 12 to 7 minutes, but visitors also fell by half. The decrease is real, yet lower demand is an alternative explanation. Compare periods with similar visitor volume or a comparable site without the extra entrance. Do not conclude that the entrance had no effect; the current evidence cannot isolate its effect.

Apply the skill: evaluate a fresh argument

A town newsletter proposes extending the Saturday repair cafe by two hours. It reports that 24 people left before receiving help last month and that volunteers are willing to stay longer. A resident objects that staff must also pay for the building's extra hours. The newsletter replies that donated materials make repairs inexpensive but gives no information about building costs.

Apply the same evidence: another decision

Quick Reference

Claim → evidence → reasoning → assumptions. Test relevance, sufficiency, independence and scope. Strengthen the actual weak link; separate observed change from demonstrated cause.