Applying Experimental Designs
- In this skill you are handed a described study and asked to pick, or name, the design that suits it.
- You base the choice on whether the same people or different people should meet the conditions.
- You only ever need the two main designs: independent measures and repeated measures.
- Always ask first whether doing both conditions would spoil one of them.
- In the exam you must name a design and give a reason it fits, never the name on its own.
A Worked Scenario
- A psychologist wants to know whether background chatter makes people worse at spotting spelling mistakes.
- The independent variable is the setting: a silent room or a recording of Greggs at lunchtime played through speakers.
- The dependent variable is how many planted spelling errors each person finds on a one-page article.
Weighing Up the Two Designs
- Repeated measures
- The same people proofread once in silence and once in noise, but by the second page they have had practice and may be tiring.
- Independent measures
- Different people work in each setting, so nobody carries practice across, though sharp-eyed proofreaders might land in one group by chance unless you allocate people randomly.
- Do not pick repeated measures here without spotting that proofreading the first page teaches people what to look for.
- Do not stop at naming a design: the marks are for the reason it suits this study.
Choosing the Design
- Because finding errors on the first page teaches people what to look for, doing both settings would spoil the second one.
- So different people should work in each setting, which points to independent measures.
- Use random allocation to decide who goes in each group, so naturally sharper readers are spread across both settings rather than bunched in one.
- Put the names of all 40 sixth-formers into a hat.
- Draw them out one at a time, sending the first to silence, the second to noise, and so on.
- Chance now decides the groups, so any difference in errors found is harder to blame on one group simply being better readers.
A Step-by-Step Routine
- Ask whether the same people could safely do every condition without one spoiling another.
- If they can, repeated measures works; if they cannot, use independent measures with random allocation.
- Name your design and back it with one reason drawn from the study in front of you.
- What is the first question to ask when choosing a design for a study?
- Why would repeated measures be a poor choice for the proofreading study?
- How does random allocation help an independent measures design?
- Besides naming a design, what must a full answer include?