Applying Repeated Measures Designs
- In this skill you either recognise a repeated measures design in a described study or plan one yourself.
- The giveaway is that the same people take part in every condition.
- Once you have spotted it, your job is to decide whether order effects need controlling.
- Same participants in all conditions means repeated measures.
- If doing the first condition could change performance on the next, plan counterbalancing.
A Worked Scenario
- You want to know whether people react faster with their dominant hand or their other hand.
- A partner drops a ruler between their fingers and you record how far it falls before they catch it.
- Comparing each person's two hands means the same people do both conditions, so this is repeated measures.
- The independent variable is which hand is used, and the dependent variable is the catch distance in centimetres.
Deciding Whether You Need Counterbalancing
- Ask whether doing the first hand changes how someone does the second.
- It does, because people get the hang of the timing, so a practice effect is likely.
- So you counterbalance the order rather than letting everyone use the dominant hand first.
- Half your class start with the dominant hand, then switch to the other hand.
- The other half start with the non-dominant hand, then switch to the dominant one.
- Now any practice from the first go is shared evenly between the two hands.
- A rest between the two goes also keeps fatigue from piling onto the second catch.
A Step-by-Step Routine
- Confirm from the scenario that the same people do all the conditions.
- Judge whether the first condition could affect the next through practice or fatigue.
- If it could, name counterbalancing and state the exact split (half in each order); the method scores separately from the term.
- What tells you a study uses repeated measures?
- Why is counterbalancing needed in the ruler-drop study?
- How would you counterbalance a study with two conditions?
- When might a repeated measures study not need counterbalancing at all?