Weigh Evidence by Its Strength

You can now name the four kinds of evidence. But naming a piece of evidence only tells you what it is, not how much weight it can bear. A shaky measurement and a careful one are both “data,” yet you would never lean on them equally. This chapter gives you a simple way to judge how strong any piece of evidence is, and to compare pieces fairly even when they come in different kinds.

See Why the Kind of Evidence Is Not Enough

It is tempting to rank the kinds of evidence once and for all: data beats a story, an expert beats an observation, and so on. That instinct feels tidy, and it will lead you straight into weak claims. A carefully kept logbook can be worth far more than a number someone jotted down in a hurry. A single clear observation can outweigh a big pile of figures that miss the point. The label on a piece of evidence, whether data, observation, or testimony, does not fix its worth. What fixes its worth is how strong that particular piece is for the particular claim in front of you. So before you compare your evidence, you need a way to measure strength itself.

Judge Any Piece of Evidence on Four Questions

Whatever kind of evidence you are holding, you can size up its strength by asking four plain questions. None of them requires any special training, just the willingness to ask.

Relevance is the first: does this evidence actually bear on the claim you are making? A gym’s glowing membership numbers say nothing about whether its classes are any good. Strong-looking evidence about the wrong thing is worthless, no matter how solid it is on its own terms.

Reliability is the second: can you trust where the evidence came from? A price you read off an official receipt is more reliable than one you half-remember. A figure from a source with something to sell is worth less than the same figure from a neutral one. Ask who produced the evidence, and whether they had any reason to shade it.

Coverage is the third: how much does the evidence actually take in? One person’s experience of a bus route is a thin slice; two weeks of daily timings is a much fuller picture. Evidence that rests on a single case can be true and still mislead you, because one case may not be typical.

Verifiability is the fourth: could someone else check it? Evidence you can point to, such as a record, a receipt, or a source others can consult, is stronger than evidence that lives only in one person’s memory. If no one can ever check it, you are being asked to take it purely on trust.

Compare Kinds on a Spectrum, Not a Ladder

Here is the mistake to avoid: imagining the four kinds of evidence as rungs on a fixed ladder, with data at the top and testimony at the bottom. Strength does not work like that. It sits on a spectrum, and any kind of evidence can land anywhere along it depending on how it scores on the four questions.

A strong direct observation can easily beat weak data. If you watch a shelf collapse the moment a heavy box goes on it, that firsthand sighting tells you more than a vague spreadsheet of “shelf incidents” compiled by someone who never saw one. Equally, careful data can beat a confident expert who is guessing outside their field. The point is not which kind wins in general, because no kind wins in general. The point is which specific piece is stronger here, judged on relevance, reliability, coverage, and verifiability.

Don’t Mistake Strong-Sounding for Strong

Some evidence sounds strong without being strong, and telling the two apart is a large part of weighing well.

Not on its own. Vividness, confidence, and a numerical sheen all make evidence feel weightier than it is. A dramatic, detailed story sticks in the mind, but one striking case still has almost no coverage. A speaker who sounds completely certain has not, by sounding certain, made their evidence any more reliable. And a precise-looking figure, such as “sales dropped 37.2 percent,” carries no weight at all if it comes from a source you cannot trust or measures the wrong thing. When a piece of evidence impresses you, pause and run it through the four questions anyway. Often the most memorable evidence is not the strongest, and the strongest evidence is quietly unremarkable.

Weigh Two Pieces Side by Side

Let’s put this to work on one small decision. Imagine your favorite neighborhood café seems slower than it used to be, and you want to know whether that is really true before you grumble or take your coffee elsewhere. You have two pieces of evidence.

The first is a furious online review posted last week: “Waited FOREVER, worst service in town, never again!” It is vivid and confident. But weigh it: its relevance is only partial (it is about service, but “forever” tells you nothing precise), its reliability is low (an anonymous, angry stranger), its coverage is tiny (one visit on one bad day), and it is essentially unverifiable. Strong-sounding, and genuinely weak.

The second is your own notes: across six visits over two weeks, you quietly timed how long your order took, and the wait averaged about nine minutes, up from the two or three you remember. Weigh it: highly relevant (it measures exactly the thing in question), reasonably reliable (you recorded it yourself as it happened), decent coverage (six visits, not one), and verifiable (you could easily time a few more). Modest and unflashy, but genuinely stronger.

Notice what happened. The louder, more emotional piece lost, and the calm, ordinary one won, not because of what kind of evidence each was, but because of how each scored on the four questions. That is exactly what weighing evidence by its strength looks like in practice.

Let’s Recap!

• The kind of evidence does not fix its strength: a weak piece of any kind can mislead you, and a strong piece of any kind can carry a claim;

• You can judge any piece of evidence by asking four questions: is it relevant, is it reliable, how much does it cover, and can it be verified;

• Strength sits on a spectrum, not a fixed ladder, so a strong observation can outweigh weak data and “data always wins” is a trap;

• Vivid, confident, or numerical evidence can sound strong without being strong, so run impressive evidence through the four questions anyway;

• When you compare two pieces, the stronger one is whichever scores better on relevance, reliability, coverage, and verifiability, not whichever feels more forceful.

You can now weigh how strong a piece of evidence is. But some things that show up in an argument are not evidence at all. The next chapter shows you how to tell a genuine claim apart from an opinion, an assertion, and a one-off anecdote.