Experimental designs

What are the different types of experimental designs?

• Independent groups

• Repeated measures

• Matched pairs

What is independent groups?

Two separate groups experience two different conditions (all participants experience one level of IV), their performances then compared

How are participants placed in each condition?

Randomly allocated to avoid researcher bias

What kind of data does independent groups produce?

Unrelated data: the data in one condition cannot be paired with the data of the other condition

What is repeated measures?

All participants experience both conditions, the two mean scores from both conditions compared

How do repeated measures counteract order effects?

Counterbalancing: half the participants take part in condition A → B, half B → A

What is matched pairs?

Participants are paired together based on a similar variable relevant to the experiment

What is the purpose of matched pairs?

To control confounding variables

What are the limitations of independent groups?

• If more participants with a particular characteristic are randomly assigned to one of the groups, this can influence the measurement of the DV → this would be a confounding variable

• Less economical than repeated measures as each participant contributes a single result only → repeated measures would produce double this

• Less time efficient

What are the strengths of independent groups?

• Order effects are not a problem like they are for repeated measures

• Participants are less likely to guess the aims

What are the limitations of repeated measures?

• Though counterbalancing attempts to fix the issue, order effects can occur because of fatigue or boredom.

• Order effects may improve performance due to practice, or deteriorate due to boredom or fatigue.

• More likely for participants to work out the aim of the study as they experience all conditions.

What are the strengths of repeated measures?

• Participant variables are controlled → higher validity

• More time efficient

• More economical - fewer participants are needed

What are the limitations of matched pairs?

• Time-consuming

• Expensive - pre-testing - less economical

• Participants can never be matched exactly - small participant variables

What are the strengths of matched pairs?

Order effects and demand characteristics are not problems as participants only take part in a single condition