Experimental Design

When conducting an experiment, psychologists must decide how to assign participants to conditions. This is called the experimental design. The choice of design affects how the results can be interpreted and what conclusions can be drawn. There are three main experimental designs used in psychology.

flowchart TD
    A[Experimental Design] --> B["Independent Groups<br/>different participants per condition"]
    A --> C["Repeated Measures<br/>same participants in all conditions"]
    A --> D["Matched Pairs<br/>different but matched on key variables"]
    B --> B1[No order effects]
    B --> B2["Risk: participant variables<br/>fix: random allocation"]
    C --> C1[Controls participant variables]
    C --> C2["Risk: order effects<br/>fix: counterbalancing"]
    D --> D1["Reduces participant variables<br/>+ no order effects"]
    D --> D2[Time-consuming to match]

---

The Three Experimental Designs

1. Independent Groups Design

In an independent groups design, different participants are used in each condition of the experiment. Each participant takes part in only one condition.

Example: Testing whether background music affects memory:

• Group A (condition 1): learns a word list in silence
• Group B (condition 2): learns a word list with background music
• Different people in each group

Advantage	Disadvantage
No order effects (boredom, fatigue, practice) because each participant only does one condition	Participant variables — differences between people in each group may affect results (e.g. one group might have naturally better memories)
Participants are less likely to guess the aim of the study	Requires more participants (twice as many as repeated measures)

How to reduce the disadvantage: Use random allocation — randomly assigning participants to conditions helps ensure that participant variables are evenly distributed.

2. Repeated Measures Design

In a repeated measures design, the same participants take part in all conditions of the experiment.

Example: Testing whether background music affects memory:

• The same participants learn a word list in silence AND learn a different word list with background music
• Their scores are compared across both conditions

Advantage	Disadvantage
Controls for participant variables — the same people are in both conditions, so individual differences cannot explain any difference in results	Order effects — doing one condition first may affect performance in the second (e.g. practice effect, fatigue, boredom)
Requires fewer participants	Participants may guess the aim of the study (demand characteristics)

How to reduce the disadvantage: Use counterbalancing — half the participants do condition A first, then B; the other half do B first, then A. This spreads order effects evenly across conditions.

3. Matched Pairs Design

In a matched pairs design, different participants are used in each condition (like independent groups), but participants are matched on key variables that might affect the results (e.g. age, gender, IQ, prior experience).

Example: Testing whether background music affects memory:

• Participants are paired based on their memory ability (measured by a pre-test)
• One person from each pair goes into condition A; the other goes into condition B

Advantage	Disadvantage
Reduces participant variables (because participants are matched on key characteristics)	Time-consuming and difficult to match participants accurately
No order effects (each participant only does one condition)	Requires more participants than repeated measures
	Impossible to match on all relevant variables

---

Comparing Designs

Feature	Independent Groups	Repeated Measures	Matched Pairs
Participants	Different in each condition	Same in both conditions	Different but matched
Order effects?	No	Yes (use counterbalancing)	No
Participant variables?	Yes (use random allocation)	No	Reduced (but not eliminated)
Number of participants	More needed	Fewer needed	More needed
Time/effort	Moderate	Moderate	High (matching)

Exam Tip: When choosing or evaluating a design, always explain both the advantage and how the disadvantage can be addressed. For example: "A repeated measures design controls for participant variables, but order effects may occur. This can be addressed by using counterbalancing."

---

Random Allocation

Random allocation means assigning participants to conditions using a random method (e.g. pulling names from a hat, using a random number generator). This is used in independent groups and matched pairs designs to reduce the effects of participant variables.

Random allocation is not the same as random sampling:

• Random allocation = how participants are assigned to conditions
• Random sampling = how participants are selected from the population

---

Key Points

• Three experimental designs: independent groups, repeated measures, matched pairs.
• Independent groups: different participants in each condition; risk of participant variables.
• Repeated measures: same participants in all conditions; risk of order effects.
• Matched pairs: different participants matched on key variables; time-consuming.
• Random allocation reduces participant variables in independent groups designs.
• Counterbalancing reduces order effects in repeated measures designs.

---

Why Experimental Design Matters

The experimental design is not just a technical detail — it shapes what conclusions the study can support. An independent groups study of a memory drug cannot tell you whether the same person's memory improves over time with repeated use; only a repeated measures design can answer that question. Similarly, a repeated measures study of anxiety cannot tell you whether people naturally differ in anxiety levels; an independent groups or matched pairs design is needed to compare groups.

Experienced researchers choose the design based on the research question, practical constraints, and the risks of participant variables, order effects and demand characteristics. Skilled examinees do the same when answering design questions in the AQA paper.

---

Matching Participants in Detail

Matched pairs is the most labour-intensive design because each participant must be paired with another on one or more variables likely to influence the DV.

Typical matching variables include:

• Age and gender (most common baseline)
• Pre-test score on the DV (e.g. memory pre-test for a memory study)
• IQ or academic attainment
• Relevant personality traits (e.g. anxiety scores for a stress study)