Experimental Design (AQA A-Level Psychology): Revision Notes
Experimental Design
Experimental design refers to how researchers organise the testing of participants in relation to different experimental conditions. To determine whether an independent variable (IV) affects the dependent variable (DV), researchers need comparison conditions - different levels of the IV to compare against each other.
This creates three main approaches to experimental design, each with distinct advantages and limitations that researchers must consider when planning their studies.
Types of experimental design
Independent groups design
An independent groups design involves using two separate groups of participants, where each group experiences only one experimental condition. This design ensures that each participant contributes data from just one condition, meaning researchers compare the performance of different people rather than the same people under different circumstances.
Worked Example: Energy Drink Study
In a study testing the effects of an energy drink on verbal performance:
- Group 1: Consumes the energy drink (experimental condition)
- Group 2: Drinks water (control condition)
- The performance of these two groups is then compared
Each participant only experiences one condition, so the researcher compares different people's performance.
The key principle here is that each participant provides data from only one experimental condition, making this a between-subjects comparison.
Repeated measures design
A repeated measures design involves all participants experiencing every condition of the experiment. This approach guarantees that researchers are comparing 'like with like' since the same participants provide data for both conditions. However, it introduces potential complications related to the order in which participants experience the conditions.
Worked Example: Energy Drink Study (Repeated Measures)
Using the same energy drink study:
- Each participant is tested after consuming the energy drink
- The same participants are later tested again after drinking water
- This creates two sets of data from the same individuals
- Researchers compare each person's performance across both conditions
This design creates a within-subjects comparison, where each participant serves as their own control, eliminating individual differences between conditions.
Matched pairs design
Matched pairs design represents a compromise between the previous two approaches. Participants are first paired based on variables that might influence the dependent variable - for instance, pairing participants with similar IQ scores in a memory study. One member of each pair is then assigned to one experimental condition whilst their partner is assigned to the other condition.
This design attempts to control for individual differences (participant variables) whilst avoiding some of the problems associated with repeated measures designs. However, perfect matching is impossible, and the process can be time-consuming and expensive.
The matching process typically involves pre-testing participants on relevant variables or using existing data (such as age, gender, or ability scores) to create comparable pairs.
Control methods
Random allocation
Random allocation is used specifically with independent groups designs to control for participant variables. This method ensures that each participant has an equal chance of being placed in any experimental condition, helping to distribute individual differences evenly across groups.
Without random allocation, researchers might inadvertently create groups that differ systematically, making it impossible to determine whether results reflect the IV's effect or pre-existing group differences.
Counterbalancing
Counterbalancing addresses order effects in repeated measures designs. This involves having half the participants experience conditions in one order (e.g., condition A then condition B) whilst the other half experience them in the opposite order (condition B then condition A).
This technique helps control for any advantages or disadvantages that might arise from experiencing conditions in a particular sequence, such as practice effects or fatigue.
Evaluation of experimental designs
Independent groups design
Independent groups designs offer several advantages, particularly in avoiding complications related to participants experiencing multiple conditions. However, they also present significant challenges in terms of controlling for individual differences between groups.
Strengths:
- Order effects are not problematic since participants only experience one condition
- Participants are less likely to guess the study's aims, reducing demand characteristics
- No risk of practice or fatigue effects affecting performance
Limitations:
- Participant variables pose the main threat - differences between groups might reflect individual differences rather than the IV's effect
- Less economical as twice as many participants are needed to generate equivalent amounts of data
- Requires random allocation to be effective, which may not always eliminate group differences
Repeated measures design
Repeated measures designs excel at controlling for individual differences but introduce their own set of methodological challenges that researchers must carefully manage.
Strengths:
- Participant variables are controlled since the same people provide data for all conditions
- More economical - fewer participants needed to generate sufficient data
- Greater sensitivity to detect genuine effects of the IV
Limitations:
- Order effects can be problematic - the sequence of conditions may influence performance through practice, fatigue, or carry-over effects
- Demand characteristics are more likely as participants experience all conditions and may deduce the study's purpose
- Counterbalancing is essential but may not completely eliminate order effects
- Some studies cannot use this design (e.g., when conditions involve permanent changes)
Matched pairs design
Matched pairs designs attempt to combine the benefits of both other approaches, though they bring their own practical and theoretical limitations that researchers must consider.
Strengths:
- Provides some control over participant variables through the matching process
- Avoids order effects since participants only experience one condition
- Reduces demand characteristics compared to repeated measures
Limitations:
- Perfect matching is impossible - even identical twins have important differences that might affect the DV
- Time-consuming and expensive, particularly when pre-testing is required for effective matching
- Still vulnerable to participant variables, though less so than independent groups designs
- May require large participant pools to find suitable matches
Key Points to Remember:
- Independent groups design uses separate groups for each condition - control participant variables through random allocation
- Repeated measures design uses the same participants in all conditions - control order effects through counterbalancing
- Matched pairs design matches participants on relevant variables before splitting them between conditions
- Each design has specific control methods: random allocation for independent groups, counterbalancing for repeated measures
- Consider practical factors like time, cost, and participant availability when selecting an experimental design