Experimental Design | Psychology for A Level - Year 13 PDF Download

Independent Groups Design & Random Allocation

• There are three primary experimental designs:
  • Independent Groups Design
  • Repeated Measures Design
  • Matched Pairs Design
• In an independent groups design, each participant is exposed to only one level of the independent variable (IV). 
  For example,
  • Participant A learns a poem while music plays (condition 1).
  • Participant B learns the same poem in a quiet environment (condition 2).
• This design produces unrelated data, meaning each group generates a distinct dataset because separate groups are used.
  • The performance of the group in condition 1 (e.g., recalling a poem learned with music) is compared to the performance of the group in condition 2 (e.g., recalling the same poem learned in silence). The dependent variable (DV) is measured as the number of words accurately recalled from the poem within a 5-minute period. Each participant provides one score based on their single condition.
• To assign participants to conditions, random allocation is used to prevent researcher bias.
   For Example, 

1. Each participant’s name is placed in a container.
2. The first name drawn is assigned to condition 1, the second to condition 2, and so forth until all participants are assigned.
3. For larger samples, software may be used to randomize allocation.

• This randomization ensures fairness and minimizes bias in assigning participants to conditions.
  In experiments, researchers manipulate the IV to observe its impact on the DV while controlling for other variables.

Evaluation of Independent Groups Design & Random Allocation

Strengths

• Reduced Demand Characteristics: Since participants engage in only one condition, they are less likely to discern the study’s purpose and alter their behavior, enhancing the study’s internal validity.
• Elimination of Order Effects: Participants experience only one condition, avoiding fatigue, boredom, or excessive practice, which strengthens the validity of the results.

Limitations

• Participant Variables: Random allocation may inadvertently place participants with similar traits in one condition, skewing results and reducing the validity of the IV’s effect on the DV.
• Need for More Participants: This design requires a larger sample, which can pose logistical challenges if participant availability is limited. Smaller group sizes may reduce the reliability of findings due to limited data.

Repeated Measures Design & Counterbalancing

• In a repeated measures design, each participant experiences all levels of the IV. 
  For example,
  • Participant A learns a poem with music (condition 1) and a different poem in silence (condition 2).
  • This design generates related data, as each participant’s performance in condition 1 is compared to their performance in condition 2, with participants serving as their own control group.
• For Example, 
  • Participant A’s ability to recall a poem learned with music (condition 1) is compared to their ability to recall a different poem learned in silence (condition 2), with the DV being the number of words correctly recalled in 5 minutes.
• Repeated measures designs may lead to order effects, including:
  • Fatigue: Participating in multiple conditions may tire participants, potentially reducing performance in later conditions.
  • Boredom: Prolonged participation may cause participants to lose interest, leading to decreased effort.
  • Practice: Performing similar tasks in multiple conditions may improve performance in later conditions.
• To mitigate order effects, researchers use counterbalancing:
  • The participant group is divided into two (e.g., 20 participants in each group).
  • One half experiences condition A then condition B, while the other half experiences condition B then condition A.

Evaluation of Repeated Measures Design & Counterbalancing

• Control of Participant Variables: Since participants serve as their own controls, individual differences are minimized, increasing internal validity.
• Fewer Participants Needed: Each participant provides data for all conditions, reducing the need for a large sample and simplifying recruitment.

Limitations

• Demand Characteristics: Exposure to all conditions may allow participants to guess the study’s aim, potentially altering their behavior and reducing internal validity.
• Order Effects: If not properly managed, fatigue, boredom, or practice effects may influence results, making it unclear whether the IV or external factors caused the observed outcomes.

Matched Pairs Design

• ​In a matched pairs design, participants are paired based on specific characteristics relevant to the study, such as:
  • Age
  • Gender
  • Ethnicity
  • IQ
  • Aggression
• For Example,
  • ​Maguire et al. (2000), taxi drivers were matched with controls based on age, gender, and handedness (right-handedness). 
  • Matching ensures that conditions are balanced, preventing an overrepresentation of certain traits (e.g., more males or older individuals in one condition).
• Matched participants are then randomly assigned to one condition each, producing related data since each participant is compared to their matched counterpart. 
• For Example, in a study on social learning of aggression:
  • Participant A, with an aggression score of 10, is matched with Participant B, who also scores 10.
  • Participant A is assigned to condition 1, and Participant B to condition 2.
    • This matching minimizes aggression as a confounding variable, ensuring differences in outcomes are due to the IV.
• Identical (monozygotic) twins are often used in matched pairs designs due to their identical DNA and similar upbringing, with one twin assigned to the experimental condition and the other to the control.

Evaluation of Matched Pairs Design

Strengths

• Control of Individual Differences: Matching participants reduces the impact of individual differences, increasing the reliability of results by comparing similar individuals.
• Reduced Demand Characteristics: Since each participant experiences only one condition, they are less likely to guess the study’s aim, enhancing validity.

Limitations

• Matching Challenges: Finding participants who match on all relevant criteria is time-consuming and difficult, especially if unmatched traits could influence results, reducing reliability.
• Participant Dropout: If one participant withdraws, replacing them with a closely matched individual is challenging, potentially delaying research or risking funding if timelines are involved.