Experimental designs

Experimental designs

What are the types of experimental designs?

• Repeated measures

• Independent groups

Experimental designs

1. What’s independent groups?

2. How is it done?

3. What data does it produce?

1. Different Ps complete each condition/variation

2. Randomly allocate Ps to each condition (avoid researcher bias) + the overall findings of each condition = compared

3. Produces unrelated data → the individual data points in one condition cannot be paired with any of the data points in the other condition

Experimental designs

What are the advantages of independent groups?

• Less likely to work out aim vs repeated measures bcs only take part in one condition SO reduces chance of participant variables

• No order effects bcs take part in one condition

Experimental designs

What are issues w/ independent groups design?

• Participant variables → if more Ps w/ a particular characteristic are randomly assigned to one group (eg: age) can influence results + show difference in conditions that isn’t really there (extraneous variable)

• Need more participants vs repeated measures to collect same amount of data

Experimental designs

How can you attempt to control the effects of participant variables in independent groups design?

How is it conducted?

• Matched pairs design → Ps are matched w/ Ps in other conditions based on a relevant characteristic (eg: aggression)

• Conducted by:

  1. Ranking Ps in order of most to least of a relevant characteristic measured before the study (eg: aggression, height, age)

  2. The top 2 = paired, then next 2, and so on until whole sample is paired

  3. Ps from each pair = randomly assigned into separate conditions

Experimental designs

What are positives of matched pairs?

• Controls participant variables (vs independent groups) → should have same overall mean of characteristic in each condition (eg: same mean age etc)

• Produces related data (vs unrelated in independent groups) SO each P’s score/data point in one condition can be paired w/ their data point in the other condition

• No order effects bcs Ps only do one condition

Experimental designs

What are issues with matched pairs?

• Takes longer than other designs

• Need more participants vs repeated measures to collect same amount of data

• May still be participant variables bcs Ps are SIMILAR BUT NOT THE SAME

Experimental designs

1. What’s repeated measures?

2. What type of data is produced?

1. All participants complete all conditions/variations of the IV

2. Related data → each P’s score/data point in one condition can be paired w/ their data point in the other condition

Experimental designs

What are advantages of repeated measures?

• Requires less Ps vs independent groups to produce same amount of data

• No participant variables effecting results of each conditions bcs same Ps complete both conditions

Experimental designs

What are issues w/ repeated measures?

• Order effects → taking part in first condition influences performance in second condition (eg: worse bcs fatigue/boredom OR improved bcs of practices)

• Demand characteristics → Ps = more likely to figure out aim + alter behaviour to prove/disprove

Experimental designs

How can you try to control (not eliminate) order effects in repeated measures designs?

Counter-balancing:

• ABBA format

• Half Ps do condition A first + condition B second

• Half Ps do condition B first + condition A second

• SHOULD cancel out order effects from other half