Evaluating Evidence and Reasoning

Earlier work with argument focused on finding evidence and matching it to a claim. Evaluating evidence and reasoning asks a harder question: even after evidence is present, does it actually justify the conclusion an author draws from it? A statistic can be accurate and still fail to support the claim it is attached to. An expert can be credible and still be quoted on a point outside their expertise. This lesson trains you to interrogate the logical bridge between what an author presents and what an author concludes, a skill closely related to Supporting Arguments With Evidence.

Every argument rests on three connected parts. The claim is the position being defended. The evidence is the fact, data, or testimony offered in support. The reasoning, sometimes called the warrant, is the unstated or explicit logic that explains why this evidence should lead a reader to accept this claim. Weak arguments often have solid evidence but a broken or missing warrant, the very link a careful reader must test.

To evaluate evidence rigorously, ask three questions in sequence. Is the evidence relevant, meaning does it actually bear on this specific claim rather than a nearby but different one? Is the evidence sufficient, meaning is there enough of it, and of the right kind, to justify the scope of the claim being made? Is the evidence accurate, meaning has it been represented completely, or has a detail been trimmed or a statistic taken out of context to make it look more decisive than it is? An argument can pass one of these tests and fail another; a single accurate statistic, for instance, is rarely sufficient on its own to support a sweeping generalization.

Once evidence has been tested, examine the reasoning that connects it to the claim. Look for unstated assumptions: premises the author needs to be true but never actually defends. Look for hasty generalizations, where a pattern observed in a small or unrepresentative sample is presented as universal. Look for false causation, where two events that merely occurred together are treated as cause and effect. Look for cherry-picking, where an author selects only the evidence that fits a preferred conclusion while omitting evidence that complicates it. Naming these weaknesses precisely, rather than simply calling an argument "weak," is what separates careful evaluation from vague criticism, a distinction developed further in Analyzing Evidence For Claims.

A rigorous evaluation also considers what the author leaves out. Could the same evidence support a different conclusion? Has the author acknowledged and responded to the strongest objection to their claim, or only to a weaker version of it, a move known as addressing a straw man? Arguments that anticipate objections and explain why the evidence still favors their claim after those objections are considered are logically stronger than arguments that simply ignore alternative explanations. This comparative judgment is the core skill practiced in Assessing Claims And Evidence.

Consider this passage: "Sales of the new product rose sharply the month after the company's advertising campaign launched, proving the campaign caused the increase." Applying the framework above, the evidence (a sales increase) is relevant and likely accurate, but the reasoning contains a gap: correlation in time is treated as proof of causation without ruling out other explanations, such as a seasonal demand spike or a competitor's product being discontinued that same month. A strong evaluation would name this gap explicitly: the argument needs additional evidence, such as a comparison against a period without the campaign, before the causal claim is justified.

Warrant: The underlying logic, stated or unstated, that explains why given evidence supports a particular claim.

Sufficiency: Whether the amount and kind of evidence presented is adequate to justify the scope of the claim.

Unstated Assumption: A premise an argument depends on that the author never explicitly states or defends.

Hasty Generalization: Drawing a broad conclusion from a sample that is too small or unrepresentative to support it.

False Causation: Mistaking a sequence or correlation between two events for a proven cause-and-effect relationship.

Cherry-Picking: Selecting only the evidence that supports a preferred conclusion while omitting contradicting evidence.

Straw Man: Responding to a weakened or distorted version of an opposing argument instead of its strongest form.

Related lessons

• Evidence That Proves Claims