Study Notes on Experimental Design in Psychology
Design of Experiments
Overview of Chapters 5-8
- Chapters focus on the design of experiments in psychology.
Chapter 5: Essential Features of Experimental Research
Key Components of Experiments
- Independent Variables: Factors varied in the experiment.
- Dependent Variables: Outcomes measured in the experiment.
- Extraneous Variables: Other factors controlled to isolate the effect of the independent variables.
Objectives of Chapter 5
- Understand essential features of experiments.
- Assess how design impacts validity of studies as outlined below:
- Impact of Robert Woodworth's Experimental Psychology (1938).
- Definition and examples of manipulated independent variables.
- Distinction between experimental and control groups.
- Application of John Stuart Mill's rules of inductive logic.
- Recognition and implications of confounding variables.
- Identification of independent and dependent variables in various experiments.
- Differentiation between manipulated and subject variables.
- Factors affecting statistical conclusion validity.
- Construct validity in experiment design.
- Importance of internal and external validity.
- Threats to internal validity and their implications.
- Ethical guidelines in running subject pools.
Historical Context: Woodworth’s Contribution
- Robert Sessions Woodworth's Experimental Psychology published in 1938 influenced the definition of experiments:
- Defined experiments with rigorous standards contrasting them from correlational research.
- Introduced terminology like "independent variable" and "dependent variable."
- Origin of methods established by John Stuart Mill which aid in establishing causality in experiments.
John Stuart Mill's Inductive Logic
- Method of Agreement: Identifies conditions where a factor (X) consistently leads to an outcome (Y).
- Method of Difference: If X does not occur, the outcome (Y) is absent.
- Joint Method: Combining both methods enhances confidence that a relationship exists.
Establishing Independent Variables
- Defined as the manipulated factor in an experiment.
- Must have at least two levels for comparison.
- Example: Study on caffeine effects requiring two dosage levels.
Categories of Manipulated Independent Variables
- Situational Variables: Environmental features affecting behavior.
- Example: Number of bystanders in a helping behavior study.
- Task Variables: Variations in the tasks assigned to participants.
- Instructional Variables: Different instructions given to participants.
Control Groups
- Experimental Group: Receives treatment or manipulation.
- Control Group: Does not receive treatment, serving as a baseline.
- Importance of ensuring both groups are as identical as possible to isolate effects of treatment.
Example Study: Superstition and Luck
- Experiments by Damisch, Stoberock, and Mussweiler (2010) demonstrated how beliefs in luck affect performance on tasks.
- Confounding: Occurs when extraneous variables co-vary with independent variables, affecting interpretation of results.
- Need to hold extraneous factors constant to avoid confounds.
- Real-life examples of confounding are elaborated using hypothetical studies.
Measuring Dependent Variables
- Dependent variables are the behavior outcomes measured in the study.
- Decisions regarding how dependent variables are defined affect both construct and statistical validity.
- Operational definitions must be precise for reliability.
Subject Variables
- Variables not manipulated by the researcher, existing prior to the study.
- Common in studies on group differences, e.g., gender, age.
- Conclusions about causality can only be assumed for manipulated variables.
- Example: Study comparing anxiety levels in different natural groups.
Validity of Experimental Research
- Types of validity:
- Statistical Conclusion Validity: Accuracy in statistical analyses and conclusions drawn.
- Construct Validity: Appropriateness and accuracy of operational definitions used.
- External Validity: Generalization of results to other populations, environments, times.
- Internal Validity: Methodological soundness and control from confounding variables.
Threats to Internal Validity
- Addressed through several potential pitfalls:
- History
- Maturation
- Testing
- Instrumentation
- Selection effects
Conclusion on Internal and External Validity
- Internal validity relates to methodological rigor, while external validity concerns generalizability.
- Importance of maintaining a balance between internal and external validity in research design.
Ethical Considerations in Research
- Essential to meet ethical standards when running experiments, especially concerning subject pools.
- Ensure participant rights and well-being are prioritized.