Why is it so hard to trust statistics?

Answer each question with 2–3 sentences.

1. Do you trust the numbers in news reports?
2. Have you ever been misled by a graph?
3. Are you good at maths?
4. Which statistic has surprised you recently?
5. Do you check where figures come from?
6. Do you think numbers are more convincing than stories?

Read each word and the example. Then say a sentence about numbers.

Choose the word that fits.

1. A small, unrepresentative can give a false picture of a whole population.
2. A link between two things that change together is a .
3. Showing that one thing makes another happen needs evidence of , not only of a link.
4. One extreme value that is very different from the rest is an .
5. The of women in the group rose from 30 to 45 per cent.
6. Every person who answers a questionnaire is a .
7. Readers may the figures, taking them to mean something the authors never intended.
8. Some headlines small risks to attract more readers.

Choose the right answer, then press Check answers.

1. a sample is a small group chosen to
   replace a larger oneentertain a larger onerepresent a larger one
2. an outlier is a value that is very
   different from the restsimilar to the resthard to read
3. a proportion is a
   type of pollkind of graphpart of a whole
4. a respondent is someone who
   prints a surveyanswers a surveywrites a survey
5. to exaggerate = to make something seem
   older than it issmaller than it isbigger than it is
6. a margin of error is the range in which the true value probably
   runsliessleeps

Use these phrases to describe figures and to question them.

• A third of the respondents...Третина респондентів...A third of the respondents said they had changed jobs.
• The vast majority of...Переважна більшість...The vast majority of people trust their doctor.
• Just under half of...Трохи менше половини...Just under half of the sample refused to answer.
• They were twice as likely as... to...Вони були вдвічі ймовірніше, ніж..., щоб...They were twice as likely as others to vote.
• ... accounted for 60 per cent of...... становили 60 відсотків...Women accounted for 60 per cent of respondents.
• The number of... has risenКількість... зрослаThe number of applicants has risen sharply.
• A number of people...Чимало людей...A number of people were undecided.
• Correlation does not prove causation.Кореляція не доводить причинності.As the saying goes, correlation does not prove causation.

Read the text. Hover over the dotted words for help.

1. A good poll can still have a margin of error.
   TrueFalse
2. The text says that ice cream causes drowning.
   TrueFalse
3. Wald is said to have suggested protecting the places with no bullet holes.
   TrueFalse
4. The text recommends ignoring all statistics.
   TrueFalse
5. The origin of the "lies, damned lies" saying is certain.
   TrueFalse

1. Which numbers do you tend to believe, and why?
2. Who should check statistics in the news?
3. Is it fair to say that numbers can lie?
4. How can schools teach statistical thinking?

Common mistakes: the number of takes a singular verb, a number of takes a plural verb; use twice as likely as, three times as many as, accounted for, just under. ✗ The number of respondents were 1,000 → was. ✗ A number of respondents was undecided → were. ✗ twice likely as → twice as likely as. ✗ accounted to 60% → accounted for 60%.

1. The number of respondents
   werewas
   1,000.
2. A number of respondents
   werewas
   undecided.
3. The vast majority of people
   saysays
   they trust their doctor.
4. Smokers were
   twice as likely astwice likely than
   non-smokers to fall ill.
5. Just underUnder just
   half of the sample refused to answer.
6. There were
   three times as manythree times more as many
   women as men.
7. Women
   accounted toaccounted for
   60 per cent of respondents.
8. The result is
   by longby far
   the most surprising.

Each sentence has one mistake. Type the correct words.

1. The number of respondents were 1,000. →
2. A number of respondents was undecided. →
3. Smokers were twice likely as non-smokers. →
4. Women accounted to 60 per cent. →
5. There were three times more as many men. →
6. Under just half refused to answer. →

Answer in detail. Use language for figures and the grammar of Block 3.

1. Which statistic have you seen recently, and did you trust it?
2. What proportion of your friends would answer an online survey?
3. Why might the number of respondents matter?
4. Can you give an example of correlation without causation?
5. Who would you ask to check a statistic?
6. Are people more likely to believe a number than a story?

Three people discuss statistics. Compare them. Who trusts numbers most? Who is the most suspicious? Whose view is closest to yours?

Your teacher will time you.

1. Plan for one minute, then speak for one or two minutes.
2. Part 1: describe a statistic you have seen and where it came from.
3. Part 2: say whether it can be trusted (use at least two proportion or comparison phrases).
4. Part 3: say what questions we should ask about numbers.

1. Do you read statistics in the news?
2. Do you trust opinion polls?
3. Have you ever answered a survey?
4. Do you check where a number comes from?
5. Do you find graphs easy to read?
6. Which figures matter most to you in daily life?
7. Have you ever been misled by a number?
8. Do you use data in your work or studies?

1. Who in your family is good with numbers?
2. Who do you trust to explain statistics?
3. Who in your life exaggerates?
4. Who checks the facts in your circle?
5. Who taught you how to read graphs?
6. Who should pay for polls and surveys?

1. Should news reports always show the margin of error?
2. Are polls bad for democracy?
3. Is it ever right to exaggerate to make a point?
4. Should schools teach more about statistics?
5. Do numbers make an argument stronger?