Kinds of Evidence
Facts, examples, quotations, data, and testimony are different kinds of evidence, and each is persuasive for a different kind of claim. · 12 min
When you make a claim, you owe your reader proof — but proof comes in kinds, and they are not interchangeable. A fact settles that something is so. An example makes a general point concrete. A quotation gives someone's exact words. Data show how much or how many. Testimony reports what a person saw or lived. Each kind is persuasive for a different sort of claim, and a careful writer chooses on purpose. Learn the five kinds now, and you can stop reaching for whatever evidence is nearest and start reaching for the kind the claim actually needs.
Guess before you learn
Your claim: 'Most students at your school skip breakfast.' Which kind of evidence would best support it?
The claim is about how many — a matter of quantity — so it calls for data: a count across students. One friend's story is a single example; the principal's words are a quotation about a different point; 'breakfast is important' is a general fact that never touches how many skip it. If you reached for the story, keep the pencil mark — a vivid case is memorable, but it cannot show what most students do.
9–12
3–5
Proof comes in kinds. A fact is something you can check. An example is one real case. A quotation gives someone's exact words. Data are numbers you gather. Testimony is what a person saw or lived through. Each kind fits a different job: numbers show how many, a quotation shows what was said, and a story shows what something was really like.
6–8
Evidence comes in five main kinds, and each answers a different kind of claim. A fact is a checkable statement — good for settling that something is so. An example is one instance — good for making a general claim concrete. A quotation is someone's exact words — good when the wording matters. Data are gathered numbers — good for how much, how many, or which way a trend runs. Testimony is a first-hand or expert account — good for what an experience was like. So read your claim first, then pick the kind that fits its job.
9–12
Evidence divides into five workhorse kinds, and each suits a different kind of claim. A fact is a verifiable statement, best for establishing that something is the case. An example is a particular instance, best for making an abstract claim concrete. A quotation reproduces exact words, best when the phrasing itself is the point. Data are systematic measurements, best for claims about how much, how often, or which way a trend runs. Testimony is a first-hand or expert account, best for lived experience. No kind is strongest in general; strength depends on the claim. Match the kind of evidence to the kind of claim, and even a true fact of the wrong kind leaves the point undefended.
K–2
When you say something is true, you show it. To show a plant grew, you measure it. To show a friend was kind, you tell what she did. Different claims need different kinds of proof.
Undergrad
Kinds of evidence are best distinguished by function, not format. A fact discharges a claim that something is or was the case. An example instantiates a generalization, turning an abstract claim into an inspectable case. A quotation is warranted when the exact language bears argumentative weight — in interpretation, in law, in reporting what a source committed to. Data support claims about quantity, distribution, and trend, where one case cannot generalize but a sample can. Testimony transmits lived experience or expert judgment, standing in for observation you cannot perform. The governing principle is fit: the claim's logical type fixes which kind of evidence can bear on it at all.
Postgrad
Each evidential kind carries its own mode of warrant. Facts and data ground claims in observation, data aggregated so as to license inference about populations rather than instances. Examples work by instantiation and, at scale, by enumerative induction — always exposed to the unrepresentative case. Quotation is the form proper to claims about texts and commitments, where the object of study is the language itself and paraphrase would beg the question. Testimony imports another's epistemic access, inheriting its epistemology: its force turns on the source's competence and sincerity. 'Best for a different kind of claim' therefore means the alignment of warrant-type to claim-type, not a ranking of kinds by intrinsic strength.
evidence
The facts, examples, quotations, data, and testimony a writer offers to make a claim believable. It comes in kinds, and each kind suits a different sort of claim.
Why is this true?
Why can't one kind of evidence prove every kind of claim?
Because a claim asserts a particular kind of thing — a quantity, a wording, an experience — and only evidence of the matching kind can bear on it. Numbers show how many but not what a line of a poem means; a quotation shows the wording but not how common a habit is. The claim decides which kind of proof can do the work.
To choose a kind of evidence, start from the claim, not from what you happen to have. Ask what kind of thing the claim asserts. Is it a matter of how many, how much, or which way a trend runs? Reach for data. Is it about what someone said or wrote, word for word? Use a quotation. Is it about what an experience was like, or a judgment only an expert can give? Use testimony. Does the claim need one real case to become concrete? Give an example. Does it simply need something settled as so? State the fact. Name the claim's type first, and the right kind of evidence follows from it.
Choose the kind of evidence for each claim — the steps fade as you master them
Claim: the population doubled over thirty years.
Claim: the witness said she saw it run the light.
Claim: patients are left exhausted for days.
You can now name the five kinds of evidence and reach for the one a claim calls for. But choosing the right kind is only the first test. A fact can be exactly the right kind — data for a claim about quantity — and still fail, because it is not about the precise thing you claimed. The next folio takes up that harder test: matching a specific piece of evidence to the exact claim it is supposed to support.
Practice — new ink and old, interleaved
1.Reviewing folio 5: order these scrambled sentences into one unified paragraph — claim, reason, then consequence.
- A used book costs a fraction of a new one and reads exactly the same.
- Buying secondhand is the thriftiest way to build a library.
- So a reader on a budget can own hundreds of books for the price of a few dozen new ones.
2.Match each paragraph problem to the fix it calls for.
3.'My grandfather, who fished this river for fifty years, says the catch is a tenth of what it was.' What kind of evidence is this, chiefly?
4.Match each broad topic to a question that narrows it well.
5.Reviewing folio 7: 'The medicine eased her pain. ___ it left her too drowsy to work.' Which transition fits?
6.Name the three tests you run to sharpen a thesis. Answer in one sentence.
7.Order these scrambled sentences so the logic flows: the claim, then a contrast, then the conclusion.
- So the honest verdict is mixed: real time saved, real boundaries lost.
- Remote work saves employees hours of commuting each week.
- It can, however, blur the line between the workday and home.
8.Claim: 'Ticket prices at the stadium have climbed every year for a decade.' Which kind of evidence fits?
9.Reviewing folio 6: turn this announcement into a topic sentence that makes a claim: 'This paragraph is about city parks.'