Confused by Experimental Design in Psychology? AQA Examples That Make It Clear

 

You can remember the definitions of repeated measures, independent groups and matched pairs. But when an exam question describes an investigation, you still hesitate.

Is that the design or the method? Which variable is being measured? Does random allocation mean random sampling?

Knowing the terminology and applying it are different skills. This guide helps you connect the definitions to what researchers actually do.

You will work through one fictional school-based investigation, identify its mistakes and improve it. Along the way, you will practise experimental design in Psychology, operationalisation, standardisation and counterbalancing.

These topics appear in the revised AQA A-Level Psychology Research Methods specification. The questions below are original practice material, not official AQA questions. www.aqa.org.uk

Experimental design or experimental method?

These terms describe different aspects of an investigation.

Think of it this way:

Design: Who completes which condition?

Method: What type of experiment is this?

A laboratory experiment can use repeated measures or independent groups. Identifying the method does not automatically identify the design.

For a broader introduction, explore our guide to experimental designs in Psychology. Here, we will focus on applying the concepts to a scenario.

The scenario: a school refurbishment prompts a research question

A sixth-form building is being refurbished. Contractors separate the work area from occupied classrooms and arrange appropriate dust controls. Depending on the project, specialist equipment such as a negative air scrubber may form part of a professionally designed containment arrangement. The linked US supplier illustrates this equipment; suitability for UK work requires competent assessment.

The Health and Safety Executive’s construction dust guidance explains the importance of controlling dust where it is produced. Students are not exposed to construction dust for research purposes. www.hse.gov.uk

The disruption raises a question for the Psychology class:

Does background noise affect performance on a proofreading task?

Researchers propose investigating this in a separate, suitable teaching room. They will use recorded building-work sounds at a comfortable, risk-assessed volume rather than conduct the study beside active construction.

Everything that follows is hypothetical. It illustrates Research Methods rather than reporting real findings.

Can you spot the problems?

The researcher recruits 20 volunteers.

• Ten complete a proofreading task in quiet conditions at 9am.
• Ten complete a different task with recorded noise at 3pm.
• The quiet group receives ten minutes.
• The noise group receives five minutes.
• Students choose which condition they join.

The quiet group identifies more errors. The researcher concludes:

“Background noise reduces concentration.”

Pause here. What else could explain the difference?

Possible explanations include:

• The quiet group had more time.
• One passage may have been easier.
• Time of day may have affected performance.
• Students choosing each condition may have differed in reading ability.

Noise was not the only difference between the conditions. The researcher therefore cannot confidently attribute the score difference to noise.

Independent and dependent variables: which is which?

The independent variable, or IV, is manipulated by the researcher.

In this proposed investigation, it is background sound, with two conditions: quiet and recorded building-work noise.

The dependent variable, or DV, is the measured outcome.

Here, it is proofreading performance.

Ask yourself:

What does the researcher change, and what result do they record?

Do not identify the participants, classroom or equipment as the IV simply because they appear in the scenario.

Operationalisation: make the variable precise

“Concentration” does not explain exactly what the researcher will measure.

An operationalised DV could be:

The number of the 20 deliberately inserted spelling errors correctly identified in a passage within five minutes.

Now another researcher knows what to count and how long the task lasts.

The IV also needs a precise procedure: the recording, playback equipment and sound level at the participant’s position should be specified.

Proofreading accuracy measures performance on this task. It does not measure every aspect of concentration or predict examination grades by itself.

If you struggle to make variables measurable, our guide to operationalisation in Psychology provides further examples.

Check your understanding

Question: Operationalise memory as a dependent variable.

Model answer: Measure memory as the number of words correctly recalled from a 15-word list during a two-minute written recall period.

“Give participants a memory test” is less precise because it does not specify the task or scoring measure.

The three experimental designs explained

Independent groups: different participants in each condition

Ten students complete the quiet condition. Ten different students complete the noise condition.

Each participant takes part in only one condition.

Advantage: Participants do not repeat the task across conditions, avoiding between-condition practice and fatigue effects.

Limitation: One group may contain stronger readers, affecting proofreading scores independently of noise.

Random allocation helps reduce systematic allocation bias. It does not guarantee identical groups.

Repeated measures: the same participants in every condition

All 20 students complete both quiet and noise conditions.

Advantage: Each student provides their own comparison. Stable differences in reading ability are therefore less likely to explain the difference between conditions.

Limitation: Performance on the second task may change because students have practised or become tired.

Matched pairs: matched participants divided between conditions

Researchers assess baseline proofreading performance and create pairs with similar scores. One member of each pair completes quiet conditions; the other completes noise conditions.

Advantage: The groups are made more comparable on a relevant characteristic.

Limitation: Matching takes time and does not make participants equivalent on every characteristic.

If you keep confusing the first two, revisit our comparison of independent groups and repeated measures designs.

For further reading, this university-hosted experimental design explanation uses the research terms between-subjects and within-subjects for independent groups and repeated measures. kpu.pressbooks.pub

How to explain an advantage in context

Practice question: Explain one advantage of repeated measures for this investigation.

Basic answer: It controls individual differences.

Stronger applied answer: Every student completes both sound conditions, so their baseline reading ability is present in both. This reduces the possibility that proofreading scores differ because one condition contains stronger readers.

The stronger answer connects the methodological point to the actual task.

Our guide to AO2 application in A-Level Psychology explains how to connect your knowledge to scenario details.

Counterbalancing: why everyone should not do quiet first

If everyone completes quiet conditions first, their second performance may be affected by practice or fatigue.

With two conditions, researchers could counterbalance the order:

Participants can be randomly allocated to these orders.

Counterbalancing distributes order effects across conditions. It does not stop practice or fatigue from happening.

Researchers should also use equivalent passages and balance their assignment across conditions. Otherwise, one sound condition might consistently receive the easier text.

Quick application question

Question: Explain why counterbalancing would improve this repeated-measures investigation.

Model answer: Some students complete quiet conditions first, while others complete noise conditions first. Any improvement from practising the proofreading task is therefore less likely to favour one sound condition consistently.

Random sampling and random allocation are different

Random sampling concerns selecting participants from a population.

Random allocation concerns assigning recruited participants to conditions or condition orders.

Our researcher recruited volunteers. Randomly allocating them afterwards does not turn the volunteer sample into a random sample.

Practice question: Twenty volunteers are assigned to conditions using a random-number generator. Identify the sampling method and allocation procedure.

Answer: Volunteer sampling and random allocation.

For further examples of using chance in research procedures, explore randomisation in Psychology.

Extraneous variables: what else might affect the DV?

An extraneous variable is a variable other than the IV that could affect the DV.

Your explanation should identify how it could influence this investigation.

“Keep everything the same” is too vague. Name the variable and explain an appropriate control.

Some factors are difficult to control completely. Researchers should acknowledge remaining limitations rather than claim that their procedure eliminates every alternative explanation.

Standardisation: what stays consistent?

Standardisation means keeping relevant procedural details consistent across participants and conditions.

For this investigation, standardise the instructions, time limit, text presentation and scoring rules.

The IV must still differ. Playing identical sound in both conditions would remove the intended manipulation.

A scoring rule might state:

Award one point for each correctly identified inserted error. Record incorrectly marked words separately.

This prevents the scoring system from ignoring differences between someone who identifies genuine errors accurately and someone who also marks many correct words as wrong.

Why run a pilot study?

A small preliminary trial could reveal:

• Unclear instructions.
• Passages that are too easy or difficult.
• Uncomfortable playback levels.
• An impractical time limit.
• Ambiguous scoring rules.

Researchers can then revise the procedure before the main investigation.

A pilot helps improve preparation. It does not prove that the final study will be valid.

A better version of the investigation

The researcher now proposes:

1. Recruit volunteers and obtain appropriate informed consent.
2. Use repeated measures.
3. Randomly allocate participants to the two condition orders.
4. Balance equivalent passage versions across sound conditions.
5. Standardise instructions, timing, room arrangements and scoring.
6. Use comfortable, risk-assessed sound levels and allow withdrawal.
7. Compare each participant’s proofreading scores.

The study addresses the noise question more clearly. It still requires evaluation.

For example, participants might guess the aim and alter their effort: demand characteristics.

A five-minute proofreading task might not represent an hour of ordinary revision: a question of ecological validity.

Our explanation of ecological validity in Psychology explores whether findings generalise to everyday settings and activities.

A higher average does not tell the whole story

Suppose the fictional mean scores are:

The quiet-condition mean is higher in this sample.

That alone does not establish statistical significance, explain every participant’s performance or demonstrate an effect on examination grades.

Researchers must consider individual scores, variability, appropriate statistical analysis and methodological limitations before drawing conclusions.

Experiment, correlation or case study?

The same school setting could support different investigations.

• Experiment: Manipulate recorded noise and measure proofreading performance.
• Correlation: Measure existing sound levels and task scores, then examine their association.
• Case study: Investigate one school in depth using several sources, such as interviews, observations and records.

A correlation alone cannot establish causation. A scenario involving one school is not automatically a case study.

Read our introduction to case studies in Psychology to revisit the in-depth approach.

Your final application challenge

Question: Explain why giving the quiet group ten minutes and the noise group five minutes is a problem. Suggest an improvement.

Model answer: Time allowed is an extraneous variable because it could affect the number of errors identified. The quiet group has more opportunity to find errors, so its higher score may reflect the longer duration rather than the absence of noise. Give both conditions the same five-minute limit.

Use this answer pattern when appropriate:

Name the issue → apply it to the scenario → explain its effect → suggest an improvement.

For further materials, explore our AQA Research Methods resources.

When a scenario feels confusing, return to four questions:

What changes? What is measured? Who completes each condition? What else could affect the result?

• Share Share on Facebook
• Tweet Tweet on Twitter
• Pin it Pin on Pinterest