Science for Everybody

Mark value: 2 marks

AO focus: AO1

Bullet-point answer points

• Experimental design concerns how participants are allocated or organised across the conditions of an experiment.
• A complete definition should link allocation to the different levels or conditions of the independent variable.

Full model answer

An experimental design is the way participants are allocated to the different conditions or levels of the independent variable in an experiment.

Common pitfalls and how to avoid them

Do not describe types of experiment such as laboratory or field experiments. The question is asking about participant allocation, not the setting.

Mark value: 4 marks

AO focus: AO1/AO3

Bullet-point answer points

• Strength: participants take part in only one condition, so order effects such as practice and fatigue are avoided.
• Weakness: different participants are used in each condition, so participant variables may create differences between groups.
• Random allocation can reduce systematic allocation bias but cannot guarantee identical groups.

Full model answer

One strength of an independent groups design is that it avoids order effects because each participant completes only one condition. Therefore, practice or fatigue from an earlier condition cannot affect later performance. One weakness is that the groups contain different people, so participant variables may influence the dependent variable. For example, one group may naturally have better memory than the other, reducing internal validity.

Common pitfalls and how to avoid them

A common mistake is to say independent groups uses fewer participants. It usually needs more participants than repeated measures. Also, link the weakness to participant variables rather than simply saying the groups are different.

Mark value: 4 marks

AO focus: AO2/AO3

Bullet-point answer points

• The design is repeated measures because the same participants complete both conditions.
• A problem is order effects, such as practice making reaction times faster or fatigue making them slower.
• The psychologist could counterbalance the order.
• Half could complete caffeine then caffeine-free, while the other half complete caffeine-free then caffeine.

Full model answer

The study uses a repeated measures design, so order effects may influence the results. Participants could become faster simply because they practise the reaction-time task in the first condition, or they could become tired in the second condition. The psychologist could use counterbalancing. Half the participants could complete the caffeinated condition first and the caffeine-free condition second, while the other half complete the conditions in the opposite order.

Common pitfalls and how to avoid them

Do not suggest random allocation as the main solution to order effects. Random allocation assigns participants to conditions and is most relevant to independent groups. Repeated measures order effects are controlled through counterbalancing.

Mark value: 6 marks

AO focus: AO2

Bullet-point answer points

• Identify a relevant matching variable, such as baseline memory ability.
• Pre-test all 40 participants using the same memory measure.
• Pair participants with similar baseline scores.
• Allocate one member of each pair to the background-music condition and the other to the no-music condition.
• Randomly decide which member of each pair enters each condition.
• Use the same memory procedure in both conditions apart from the IV and compare the scores.

Full model answer

The researcher could first give all 40 participants the same baseline memory test. Participants with similar memory scores could then be paired, creating 20 matched pairs. For each pair, one person would be placed in the background-music condition and the other in the no-music condition. The researcher could randomly decide which member of each pair enters each condition. Both groups would then complete the same memory task under standardised conditions, with background music as the only intended difference, and their memory scores could be compared.

Common pitfalls and how to avoid them

Avoid saying that both members of each pair complete both conditions; that would turn the study into repeated measures. Also, choose a matching variable that is relevant to the dependent variable rather than an arbitrary characteristic.

Mark value: 6 marks

AO focus: AO1/AO3

Bullet-point answer points

• Independent groups uses different participants in each condition; repeated measures uses the same participants in all conditions.
• Independent groups avoids order effects; repeated measures is vulnerable to order effects.
• Repeated measures controls participant variables; independent groups is more vulnerable to participant differences.
• Independent groups usually needs more participants; repeated measures can be more economical.
• Demand characteristics may be more likely in repeated measures.
• Random allocation is useful for independent groups, while counterbalancing is useful for repeated measures.

Full model answer

Independent groups uses different participants in each condition, whereas repeated measures uses the same participants in every condition. Independent groups therefore avoids order effects because participants do not repeat the task, but it is vulnerable to participant variables because the groups may differ in important ways. In contrast, repeated measures controls participant variables because each participant acts as their own comparison, but practice, fatigue or boredom may affect later conditions. Repeated measures also usually requires fewer participants, although demand characteristics may be stronger because participants experience every condition. Researchers can use random allocation to improve independent groups designs and counterbalancing to reduce order effects in repeated measures.

Common pitfalls and how to avoid them

Do not write two separate mini-essays without making direct comparisons. Use comparative language such as whereas, in contrast and however. Also avoid confusing random allocation with counterbalancing.

Mark value: 12 marks

AO focus: AO1/AO3

Bullet-point answer points

• Define experimental design as allocation of participants to conditions.
• Independent groups uses different participants in each condition.
• Independent groups avoids order effects but is vulnerable to participant variables and usually needs more participants.
• Repeated measures uses the same participants in all conditions.
• Repeated measures controls participant variables and uses fewer participants, but can produce order effects and demand characteristics.
• Counterbalancing can reduce order effects but may not remove carry-over effects completely.
• Matched pairs uses different participants matched on relevant characteristics.
• Matched pairs reduces participant differences without creating order effects.
• Matching is difficult, time-consuming and imperfect, and participant withdrawal may disrupt a pair.
• The best design depends on the research aim, task and likely sources of bias.

Full model answer

Experimental design concerns how participants are allocated to the conditions of an experiment. In an independent groups design, different participants complete each condition. A key strength is that order effects are avoided because no participant repeats the task. However, participant variables may create differences between the groups, so an apparent effect of the IV could actually be due to natural differences between participants. Random allocation can reduce systematic allocation bias, although it cannot guarantee identical groups.

Repeated measures uses the same participants in every condition. This controls participant variables because each person acts as their own comparison, and it often requires fewer participants. However, order effects can threaten validity. Practice may improve later performance, while fatigue or boredom may make it worse. Participants may also guess the aim more easily because they experience all conditions. Counterbalancing can spread order effects across condition orders, although carry-over effects may remain.

Matched pairs uses different participants in each condition but pairs them on important characteristics. This avoids order effects and can reduce participant variables. However, matching is time-consuming and cannot make two people identical. Researchers may also match on the wrong variables or lose the usefulness of a pair if one participant withdraws.

Overall, no design is always best. Researchers should choose the design that best controls the most important threats to validity while remaining practical and ethical for the study.

Common pitfalls and how to avoid them

Do not only describe the three designs. A discussion needs clear strengths, limitations and methodological consequences. Avoid treating counterbalancing as a perfect solution, and make the final judgement depend on the research context rather than claiming one design is universally superior.