PSYC 217 03c three types of experimental designs
Review and Overview of Experimental Designs
Goal of session: Compare and contrast three basic experimental designs essential for the course.
Designs focus on controlling for participant variables to avoid confounding issues.
Importance of Controlling Participant Variables
Participant variables can confound results, hindering confident interpretations of study outcomes.
If participant variables are not controlled, subsequent experiments may be required to determine causal relationships.
Three designs to control participant variables:
Between Subjects Design
Matched Pairs Design
Within Subjects Design
Between Subjects Design
Definition: Participants are randomly assigned to one condition only.
Notation: Denoted as "SS" for subjects/participants.
Example scenario: Assigning groups based on birth dates for a word color game.
Group 1: Birth dates from January to June.
Group 2: Birth dates from July to December.
Task: Participants verbally state the color of displayed words, ignoring the word text itself.
Explanation of potential confounds in assignment:
Group assignment based on birth dates may correlate with age and cognitive differences, influencing performance.
Outcomes represented in bar charts comparing mean scores for each group.
Matched Pairs Design
Definition: Participants paired based on a pretest variable that may influence the dependent variable (DV).
This design ensures balance in the characteristics being controlled for.
Procedures:
Measure the pretest variable (e.g., perceptual processing speed) before the main experiment.
Participants are matched and assigned to different conditions (e.g., matching vs. mismatching).
Visualization of Results:
Bar charts centered at zero, indicating performance related to pretest speeds.
Bar heights above or below zero indicate comparative speeds during the actual testing versus pretest performance.
Advantages:
Useful for small sample sizes where random assignment may fail.
Helpful in costly studies requiring participant preparation (e.g., fMRI, EEG).
Mitigates dropout issues in longitudinal studies.
Limitations:
Pretests may influence psychological experiences, potentially skewing results (e.g., practice effects during matching/mismatching tasks).
Solomon Four Group Design
Purpose: To assess the influence of a pretest on study results.
Structure: Involves four groups:
Pretest + Matching
Pretest + Mismatching
No Pretest + Matching
No Pretest + Mismatching
This design allows researchers to analyze the effect of pretest exposure on performance.
Key outcomes reveal if pretest contributes to behavior changes, guiding further research methodology decisions.
Within Subjects Design (Repeated Measures Design)
Definition: Each participant experiences all levels of the independent variable (IV).
Advantages:
Greater statistical power due to fewer resources expended (only need one group of participants).
Eliminates participant variable confounds because the same individual responds in all conditions.
Limitations:
Subject to order effects where initial exposure influences subsequent trials.
Example: Practice effects lead to improved performance in later trials unrelated to manipulation of IV.
Comparison of Experimental Designs
Repeated Measures:
Ideal for achieving higher statistical power, requiring fewer participants than between subjects.
Limitations include the challenge of controlling order effects.
Matched Pairs:
Balances individual differences among small groups but risks pretest altering performance.
Between Subjects:
Higher participant count needed; random assignment is essential to ensure validity but can lead to other potential confounds if not properly managed.
Conclusion
Decision on design depends on specific research questions, the sensitivity of participants to different tasks, and the concept being studied.
Insightful observation and feedback from participants are essential in adjusting designs for accuracy in experimental research.
Importance of ethical considerations, minimizing bias, and understanding participant behavior throughout the research process.
Thanks attendees for participation and anticipation of future learning opportunities in the course.