Lecture 10 - Ch6 part 2 student version winter 2025 [Autosaved]

Chapter 6: Research Strategies & Validity Overview

Research Strategy

• General approach and goals determined by the research question.

• 5 Quantitative Research Strategies:

  • Descriptive: Provides a “snapshot” of the current situation.

  • Correlational: Examines natural relationships between two variables.

  • Experimental: Investigates causal relationships between variables.

  • Quasi-Experimental: Similar to experiments, but lacks random assignment.

  • Nonexperimental: Examines relationships between groups or conditions without manipulation.

Validity

• Defined as “the quality or state of being true.”

• Refers to the degree to which a study accurately answers the question it intends to address.

• Indicates correctness or truthfulness of inferences made from study data.

• Issues like flawed design or poorly applicable results can undermine validity.

• Distinct from measurement validity.

External Validity

• Refers to the extent to which study results can be generalized to other populations, settings, and times.

• Considers if the same results would be observed under different conditions, with different people, or using different measures.

Threats to External Validity

• Any characteristic limiting generalization of results:

  • Example: Testing exclusively White middle-class children in an urban setting.

• Limited range of participant characteristics can threaten the external validity.

• Samples Not Representative: If groups are not representative, generalizations are weakened.

Internal Validity

• Defined as the extent to which a study provides a clear explanation of the relationship between variables.

  • Investigate if changes in the independent variable (IV) directly cause changes in the dependent variable (DV).

  • Adequate control of confounding variables crucial for establishing valid interpretations.

Threats to Internal Validity

• Factors that allow alternative explanations for results:

  • Poor design may lead to misinterpretation of results.

  • Example: Participants improve due to external factors (diet, rest) during a flu medication study.

Quality of a Research Study

• Research quality is proportional to how well it addresses internal and external validity criteria.

• Anticipate threats but cannot eliminate all.

• Being mindful of these threats aids in evaluating research quality.

Generalizing Across Participants or Subjects

• Selection Bias: Some participants are favored by the sampling method.

  • Found in college students as they differ in characteristics from other adults.

• Volunteer Bias: Differences between volunteers and non-volunteers may skew results.

  • Volunteers often show higher education levels and social motivations.

Generalizing Across Features of a Study

• Novelty effect: Participants behave differently due to the unique test situation.

• Multiple treatment interference: Previous treatments can influence current performance (e.g., fatigue effects).

• Experimenter characteristics can influence results and vary across studies.

Generalizing Across Features of Measures

• Assessment Sensitization: Participant responses may change due to prior measurement processes.

  • For instance, pre-tests raising participant awareness of behavior.

• Generosity across response measures whereby treatment may impact one measure but not another (e.g., physiological vs. behavioral).

General Threats to Internal Validity

1. Environmental Variables: Unmonitored changes in surroundings affect results; e.g., lighting, temperature.

2. Individual Differences: Characteristics of participants can misrepresent group comparisons.

3. Time-Related Variables: External events or individual changes over time can influence results in longitudinal studies.

Balancing Internal & External Validity

• High internal validity requires tight control over confounding variables, potentially creating an artificial study environment.

• Conversely, increasing external validity demands realistic conditions that might introduce new extraneous variables.

Understanding Artifacts and Variables

• Artifacts: External factors influencing measurements without being manipulated during study.

  • Include experimenter bias and participant reactivity.

• Extraneous Variables (EVs): Any variable not being studied that could influence results.

• Confounding Variables (CVs): A specific type of EV that systematically varies with the independent variable, creating alternative explanations.

Research Strategies vs. Research Designs vs. Research Procedures

• Research Strategies: Broad categories classifying research types based on questions.

• Research Designs: General frameworks that classify research based on study conduct.

• Research Procedures: Specific methodologies unique to individual studies.