Step 11 of 18 · Worked example · TEACH

← Statistics & Probability

0 of 3 done

Work through the example to CHECK to open the problems.

Reach CHECK to open the problems.Higher Level · Statistics · Step 11

Say what each design can conclude, and why.

Researchers report that students who eat breakfast score higher in tests. They then design a study where 200 volunteers are randomly assigned either a provided breakfast or no breakfast for a term. Compare the conclusions each study supports.

What each study can conclude

StudyDesignConclusion
Study 1Observe who eats breakfastAssociation only
Study 2Randomly assign breakfastCausation (for volunteers)
LimitBothVolunteers may not represent all
Randomisation spreads confounders evenly; without it, association is the ceiling.

Study 1: researchers report that students who eat breakfast score higher in tests.

Study 2: 200 volunteers are randomly assigned breakfast or no breakfast for a term.

Compare the conclusions each study supports.

Read it once, slowly. Then copy it onto your page.
  • What is going on? Two studies of one question: does breakfast affect test scores? Study 1 reports a link. It says nothing about anyone being told what to eat.
  • Study 2 is different: the volunteers were randomly assigned. Chance decided who ate breakfast.
  • What do they want? How strong a conclusion each study can support. No calculation: a judgement, with reasons.
Céim

Higher Level Maths · Statistics & Probability · Step 11 of 18

Study Design, Randomisation, Control & Ethics

Only a designed experiment with randomised allocation supports a cause-and-effect claim; an observational study can only report association.

Why it matters

The strongest available conclusion depends entirely on how the data were produced, and that is a marked part of exam answers.

The worked example

Study 1: researchers report that students who eat breakfast score higher in tests. Study 2: 200 volunteers are randomly assigned breakfast or no breakfast for a term. Compare the conclusions each study supports.

How the worked example goes

  1. Read

    • What is going on? Two studies of one question: does breakfast affect test scores? Study 1 reports a link. It says nothing about anyone being told what to eat.
    • Study 2 is different: the volunteers were randomly assigned. Chance decided who ate breakfast.
    • What do they want? How strong a conclusion each study can support. No calculation: a judgement, with reasons.
  2. Understand

    • Who decided who ate breakfast?
    • Study 1: the students chose for themselves. The researchers only recorded it. That is an observational study.
    • Study 2: the researchers assigned it, by chance. That is a randomised experiment.
  3. Reduce

    • What do I do for each study?
    • 1. Who decided? Name the type of study.
    • 2. Observed: name a confounder, then state the link.
  4. Solve

    • What does Study 1 support?
    • Observational: nobody assigned breakfast. Sleep, or support at home, is a confounder.
    • Study 1
  5. Check

    • Is it right?
    • Study 1: not assigned → link only ✓. Study 2: random, with a control → cause ✓
    • Would a bigger Study 1 fix it? No. Observing more students keeps the same confounders. A cause needs Study 2's design ✓

Key terms

sample survey · observational study · designed experiment · randomisation · control group · confounding · ethics