Chapter 6

CHAPTER 6: Research Strategies and Validity

Page 1: Introduction to Research Strategies

• Overview of research strategies and their importance in behavioral sciences.

Page 2: What is a Research Strategy?

• Definition: A research strategy refers to the general approach and goals of a research study.

• Determining Factors:

  • Type of question addressed.

  • Desired outcomes of the study.

Page 3: The Descriptive Research Strategy

• Purpose: Describes individual variables rather than relationships.

• Data Characteristics:

  • Provides a snapshot of specific characteristics of a group.

  • Data commonly presented as averages or percentages.

Page 4: Strategies Examining Relationships Between Variables

• Relationship Types:

  • Changes in one variable correlate with changes in another.

  • Relationship types include:

    • Linear

    • Curvilinear

    • Positive

    • Negative

Page 5: Understanding Variable Relationships

• The nature of how variables are connected is critical in research methodology.

Page 6: The Correlational Research Strategy

• Definition: Measures two variables for each individual.

• Example: Investigating GPA versus sleep habits (e.g., wake-up time).

• Visualization: Patterns represented in scatter plots.

• Important Note: Correlation does not imply causation.

Page 7: Correlational Study Data Example

• Context: Time spent on Facebook vs. GPA.

  • Data collected from eight college students.

  • Observation: GPA tends to decrease as Facebook usage increases.

Page 8: Comparing Two or More Sets of Scores

• Research Strategies:

  • Experimental, quasi-experimental, nonexperimental strategies.

• Key Concept: Comparison of scores where one variable differentiates groups.

Page 9: Grades Comparison Example

• High Income vs. Economically Underserved Families:

  • Score Distribution:

    • High Income: 72, 86, 81, 94, 85, 97, 89, 91 (Mean = 91.0)

    • Economically Underserved: 83, 89, 78, 80, 90, 81, 94, 89 (Mean = 81.9)

Page 10: Research Strategies Explained

• Experimental Research: Investigates cause-and-effect relationships.

• Quasi-Experimental Research: Lacks definitive control, often yielding ambiguous results.

• Nonexperimental Research: Describes relationships without explaining them.

Page 11: Data Examples Across Strategies

• Experiments: Control actual conditions (e.g., exercise levels affecting health).

• Quasi-experiments and Nonexperimental: Observational data from different groups (e.g., smoking behaviors, cholesterol scores).

Page 12: Nonexperimental vs. Correlational Research

• Goal: Both design demonstrate relationships without explaining them.

• Data Differences:

  • Correlational research utilizes one participant group for two variables.

  • Nonexperimental research compares two different participant groups for one variable.

Page 13: Research Strategy Summary: Categories

• Category 1: Descriptive research examining individual variables.

• Category 2: Correlational strategies measuring relationships between variables.

Page 14: Research Strategy Summary Continued

• Category 3: Examines relationships by comparing two or more groups (Experimental, Quasi-experimental, Nonexperimental).

Page 15: Category 1: Descriptive Research

• Purpose: Describe variables in a specific group.

• Data Collection Example: Average study hours and sleep quality among students.

Page 16: Category 2: Correlational Research

• Purpose: Describe relationships between two variables without providing explanations.

• Data Representation: Measures two variables for each participant (e.g., wake-up times and GPAs).

Page 17: Category 3: Experimental Research

• Purpose: Establish cause-and-effect regarding variable relationships.

• Data Acquisition: Manipulate one variable to measure results in another variable (e.g., exercise impact on cholesterol levels).

Page 18: Category 3: Quasi-Experimental Research

• Objective: Attempt cause-and-effect explanations without strict experimental control.

• Data Example: Comparisons of behavior changes in groups receiving or not receiving treatment.

Page 19: Category 3: Nonexperimental Research

• Goal: Describe relationships without causal explanation.

• Example: Gender differences in verbal skills without explanation of causation.

Page 20: Research Designs

• Considerations:

  • Group vs. individual focus.

  • Consistency of participants across conditions.

  • The number of variables included.

Page 21: Research Procedures

• Details Include:

  • Specific implementation steps for the research.

  • How variables are manipulated and measured.

  • Participant involvement protocol.

Page 22: Data Structures and Statistical Analysis

• Comparison Strategies: Experimental, quasi-experimental, and nonexperimental studies utilize similar statistical techniques (t-tests, ANOVA).

• Correlation: Numerical data analyzed with Pearson’s correlation, while categorical data applies chi-square tests.

Page 23: Data Structures in Descriptive Research

• Summarization Techniques:

  • Numerical data calculated via mean scores.

  • Categorical data displayed through percentage distribution.

Page 24: External Validity Explained

• Definition: The ability to generalize research findings across varied circumstances.

• Threats: Characteristics limiting generalizability (e.g., population differences).

Page 25: Types of Generalization and Associated Threats

• Types:

  • Sample to population.

  • One study to another.

  • Study outcomes to real-world scenarios.

Page 26: Internal Validity

• Focus: Factors questioning result interpretations within a study.

• Objective: Aim for clear explanations for relationships between variables; recognize alternative explanations as threats to validity.

Page 27: Validity and Research Quality

• Determining Validity: Based on internal and external validity standards.

• Threat Definition: Factors that create uncertainty regarding result accuracy and interpretations.

Page 28: Implications of Validity Threats

• Understanding Variability: Research results are not absolute truths; critical evaluation is crucial.

Page 29: Categories of Threats to External Validity

• General Threats: Three categories outlined, impacting generalizability.

Page 30: Threats to External Validity Across Participants

• Examples: Selection bias, excessive use of college student samples, volunteer bias, etc.

Page 31: Threats Related to Study Features

• Concerns: Novelty effects and experimenter influences disrupting results.

Page 32: Threats Related to Measures

• Considerations: Sensitization effects and response measure consistency affecting results.

Page 33: Extraneous Variables Overview

• Definition: Variables present aside from those being actively studied.

• Confounding Variables: Unmanaged variables impacting the relationship and posing threats to internal validity.

Page 34: Extraneous Variables Impacting Validity

• Types:

  • Environmental variables (external threats across all designs).

  • Participant variables (individual differences).

  • Time-related variables affecting repeated measures in a study.

Page 35: Common Threats to Internal Validity

• Examples: Include maturation effects, history, selection biases, drop-outs, and measurement inconsistencies.

Page 36: Maximizing Validity in Research

• Objective: Strive to enhance both internal and external validity, recognizing trade-offs.

Page 37: Threats Impacting Both Validities

• Artifacts and Bias: External influences altering measurement integrity.

• Examples: Experimenter bias and exaggerated variables compromising findings.

Page 38: Additional Threats to Both Validities

• Demand Characteristics: Participants altering behavior due to awareness of their involvement in a study, questioning validity of results.

Page 39: Construct Validity

• Definition: Assessment of research results relating back to theoretical constructs.

• Improvements: Employ manipulation checks to ensure independent variable accuracy.

Page 40: Threats to Construct Validity

• Key Factors:

  • Weak theory-link methods.

  • Ambiguous operational definitions or measurement techniques.

Page 41: Statistical Conclusion Validity

• Purpose: Distinguish between results due to chance versus actual cause-effect relationships.

• Metrics: Power and effect size as indicators of validity strength.

Page 42: Threats to Statistical Conclusion Validity

• Examples:

  • Erroneous statistical choices can compromise results.

  • Limited study power leads to undetected effects, while inaccurate effect sizes misrepresent relationships.