• Foundation
  • About 13 min
  • 5 practice questions

Evidence in arguments: examples, numbers and experts

Judge how much support an example, a survey figure or an expert's opinion gives, by asking whether it is typical, fairly chosen and relevant.

If something looks wrong, please tell us.

Every argument rests on evidence. The evidence may be an example, a survey figure, or the word of an expert. Critical reasoning does not ask whether the evidence is true. It asks whether the evidence, if true, gives good support to the conclusion. This lesson shows what to check for each kind.

Three kinds of evidence

Three kinds of evidence

  • Examples
    • Is the example typical?
    • Weak when one or two cases speak for thousands
  • Surveys and numbers
    • Is the sample large, random and similar to the group?
    • Weak when only willing people answered or the sample is tiny
  • Experts
    • Qualified on this topic and free of bias?
    • Weak when a famous person speaks outside the subject or earns from the claim
What to check for each kind of evidence
KindCheckWeak when
ExamplesIs the example typical?One or two cases are used to speak for thousands
Surveys and numbersIs the sample large, random and similar to the group?Only willing people answered, or the sample is very small
ExpertsIs the expert qualified on this topic, and free of bias?A famous person speaks outside the subject, or earns money from the claim

More things to check about numbers

  • Percentage or count. A rise from 2 to 4 cases is a rise of 100 percent but is still only two more cases.
  • What is compared. The figure must compare like with like. A fall in complaints is not proof of better service if the shop has fewer customers.
  • Who answered. If a shop asks only the customers who return a form, those who are happy may be over-represented.

Worked examples

Example 1. 'Three of my friends joined this coaching class and all passed, so this class guarantees success.' Three friends are a very small and handpicked group. Their success may have come from many causes. The example is not typical of all students, so the conclusion is weak. The word 'guarantees' is also too strong for the evidence.

Example 2. 'A well-known film actor says that this brand of cooking oil is the healthiest.' An actor is not an expert in nutrition. The actor may also be paid to say it. The evidence is not relevant to the subject, and the conclusion is weak.

Example 3. 'A shop gave a form to every customer who bought a mobile phone and received 200 replies. Of those, 180 said they were happy, so 90 percent of all customers are happy.' First check the arithmetic: 180 out of 200 is 0.9, or 90 percent. The arithmetic is correct. But only customers who chose to reply are counted. Unhappy customers may not have returned the form. The sample is self-selected, so the conclusion about all customers is doubtful.

Example 4. 'A doctor who has studied heart disease for many years advises that people over fifty should have a regular check-up.' The doctor is an expert on the right topic. Unless there is a sign of bias, the advice is good evidence for the conclusion.

Ask three things: is it typical, is it fairly chosen, and is it relevant to the claim?

— Rule to remember

Example 5. 'In a survey of 5,000 people chosen at random from the voters list, 60 percent said they prefer a morning bus service.' The sample is large and chosen without bias, so it is a good guide to the whole group. Compare Example 3, where the replies came only from people who chose to send them. The size of a sample matters, but how it was chosen matters more.

How the sample was chosen

  • Example 3: self-selected
    • 200 replies from customers who chose to send the form
    • Unhappy customers may not have replied
    • The conclusion about all customers is doubtful
  • Example 5: chosen at random
    • 5,000 people from the voters list
    • Large and chosen without bias
    • A good guide to the whole group

A note on percentages. When a question gives a percentage, ask what it is a percentage of. If 25 percent of 40 students and 25 percent of 400 students passed, the percentage is the same but the numbers are 10 and 100. Whether a rise is large depends on the base.

Key takeaways

  • Judge the strength of evidence, not just whether it sounds true.
  • Small, handpicked or self-selected samples give weak support.
  • An expert helps only on the subject of their expertise.

Check your understanding

  1. 1. 'Two of my neighbours bought shares in a company and both made a profit, so buying shares always makes a profit.' What is the main weakness?

    1. Two cases are too few to support 'always'
    2. Neighbours are not good friends
    3. Shares are bought through a broker
    4. A profit is a positive number
    Show answer

    Answer: A. Two cases cannot support a claim about all purchases, and the word 'always' is far too strong.

  2. 2. A cricketer says that a certain brand of paint is the best for house walls. Why is this weak evidence?

    1. Cricketers are paid a great deal of money for their play
    2. Cricket is a team game that needs a lot of players
    3. Being good at cricket is no sign of knowing about paint
    4. The cricketer is well known to a large number of people
    Show answer

    Answer: C. Fame in one field does not give expertise about paint.

  3. 3. A magazine asks its readers to send in a letter about their reading habits. 80 percent of the letters say they read daily. Why can this not be used for all adults?

    1. Some letters sent by post may get lost on the way
    2. Only readers who chose to write are counted in the result
    3. Magazines are printed on paper and sold in many shops
    4. Adults differ widely in the kinds of jobs that they do
    Show answer

    Answer: B. Only people who chose to write are counted, and they are likely to be keen readers.

  4. 4. Complaints against a shop fell from 40 to 20, but the number of customers fell from 4,000 to 1,000. What does this show?

    1. The two sets of figures are too different to be compared at all
    2. Service has become twice as good as it was in the earlier period
    3. The shop now has a larger number of happy customers than before
    4. Complaints per customer rose, so service may not have got better
    Show answer

    Answer: D. The rate of complaints went from 1 in 100 to 1 in 50, so the rate doubled even though the count fell.

  5. 5. A pharmacist who has worked on safe medicine storage for twenty years advises keeping medicines away from heat. How good is this evidence?

    1. Weak, because pharmacists spend most of their time in shops
    2. Weak, because twenty years is not a very long time in a career
    3. Strong, because the advice comes from an expert on this very subject
    4. Neither strong nor weak, because advice is never evidence
    Show answer

    Answer: C. The pharmacist has relevant expertise and no sign of bias, so the advice is good support.

Sign in to keep track of the lessons you finish.

Next lesson → Plans and proposals: will it work?

Privacy · Terms · Refunds · Delivery