Notes on Research Methodologies
Experimental Research
Causality
Establishing causality requires demonstrating a relationship between variables.
Covariation: Two variables must be related (e.g., when one changes, the other tends to change).
Temporal Precedence: Changes in the independent variable (IV) must occur before changes in the dependent variable (DV).
Elimination of Alternative Explanations: Rule out plausible third variables or confounds to assert true causality.
Key Features of Experimental Design
Control of Extraneous Variables: Prevent confounds to isolate cause and effect relationships.
Use of Random Assignment: Participants are randomly assigned to ensure group equality and prevent bias.
Measurement of Dependent Variable: Observe how the DV responds to changes in the IV.
Types of Designs
Single Factor Design: Study one IV with multiple levels.
Factorial Design: Study multiple IVs simultaneously to determine interaction effects.
Between and Within Subjects:
Between-Subjects Design: Different participants are used in each condition.
Within-Subjects Design: Same participants are used across all conditions.
Validity Types
Internal Validity: The extent to which the study can rule out alternative explanations for the observed effects.
External Validity: The ability to generalize findings to other contexts.
Construct Validity: Ensuring that the operational definitions of variables align with theoretical concepts.
Statistical Validity: Proper use of statistical techniques to analyze results.
Non-Experimental Designs
Definition of Non-Experimental Research: Studies where the IV is not manipulated; researchers measure variables as they naturally occur.
Types of Non-Experimental Designs
Correlational Research: Aim to find associations between variables without manipulation.
Types of correlations:
Positive Correlation: Both variables increase together.
Negative Correlation: One variable increases while the other decreases.
Zero/No Correlation: No clear relationship between variables.
Complex Correlational Methods: Include correlational matrices and partial correlations controlling for third variables.
Quasi-Experimental Designs
Definition: Designs resembling experimental research without random assignment.
Types of Quasi-Experimental Designs
One-Group Designs: Measure the same group pre and post intervention.
Nonequivalent Group Designs: Compare results between groups that are not randomly assigned.
Interrupted Time-Series Design: Set of measurements taken before and after an intervention over time.
Qualitative Research
Purpose: To understand meanings, experiences, and contexts in depth.
Methods
Interviews: One-on-one to explore in detail.
Focus Groups: Gather diverse perspectives on a topic.
Participant Observations: Engage with subjects in their natural environment.
Qualitative Data Analysis:
Use of Memos and Notes to capture insights.
Coding: Systematically classify data to form patterns.
Deductive vs. Inductive Coding: Based on existing theories vs. allowing themes to emerge from the data.
Trustworthiness
Assess findings’ applicability through transferability.
Reflexivity: Researchers acknowledge their influence on the research process.