Notes on Research Methods: Observational Methods, Correlational Studies, and Experimental Design
Overview of Key Concepts
- Genetics as a third variable: cannot infer causality from correlation; genetics can influence both variables and confound observed relationships (e.g., correlation between crime levels and ice cream consumption in large cities may be due to a third factor).
- Distinction between correlation and causation: correlational findings show relationships, not proofs of cause-and-effect.
- Observation and data collection in psychology: questions about behaviors, characteristics, beliefs, opinions, feelings, etc. Examples include age, exercise frequency, and attitudes.
- Observer-based data collection vs naturalistic observation:
- Observer report focuses on one individual or a specific target.
- Naturalistic observation records behavior as it occurs in a natural setting without intervention.
- Reactivity: if subjects know they are being observed, they may alter their behavior; goal is to observe as they actually behave.
- True experiments (experimental designs) vs observational designs:
- True experiments manipulate an independent variable (IV) and assign participants to groups to control conditions.
- Observational designs study existing conditions without manipulation, which limits causal inferences.
- Data types in research:
- Quantitative data from scales and numerical ratings (e.g., happiness on a scale, frequency counts).
- Qualitative data from interviews and open-ended responses, focusing on how rather than how much.
- Common examples used in teaching:
- Relationship between time spent on social media and outcomes.
- Attitudes toward aging and experiences of ageism.
- Classroom sociability and sharing behaviors in preschool children.
- Reporting biases in observer vs participant perspectives.
- Media portrayal vs scientific conclusions: media may claim a causal effect from correlational findings, which researchers typically caution against without experimental evidence.
- Examples of data collection methods mentioned in the transcript:
- Large-city correlations (crime level vs ice cream consumption).
- Classroom observation through a one-way mirror to assess sociability and sharing.
- Recording how often children of various ages share.
- Questionnaires with rating scales to produce quantitative data.
- Interviews to capture attitudes and experiences (qualitative method).
- Basic terminology to remember:
- Independent Variable (IV): the variable that is manipulated in an experiment.
- Dependent Variable (DV): the outcome measured.
- Experimental group: receives the IV.
- Control group: does not receive the IV.
- Random assignment: participants are allocated to groups by chance to create equivalent groups.
- Correlational study: examines relationships between variables without manipulation.
- Qualitative research: explores how people think or feel, often via interviews or thematic analysis.
- Quantitative research: collects numerical data that can be analyzed statistically.
Observational Methods
- Observer report
- Focuses on recording a single child or a specific target.
- Often used by school psychologists, teachers, or parents to observe behavior and characteristics.
- Key issue: subject should not be aware of being observed to avoid altering behavior (reactivity).
- Naturalistic observation
- Researchers observe people or children in their natural environments (e.g., classroom) without intervention.
- Example scenario: observe preschool children’s sociability in a classroom from behind a one-way mirror; measure how often they share.
- Measurement focus: frequency or quality of sociable behaviors (e.g., sharing).
- Differences between observer report and naturalistic observation
- Observer report concentrates on a particular child, often in a controlled observation setting.
- Naturalistic observation aims to capture spontaneous behavior in real-world settings.
- Reactivity and validity concerns
- If participants know they are being watched, they may change behavior, threatening ecological validity.
- The goal is to minimize observer intrusion and ensure behavior reflects actual patterns rather than manipulated responses.
True Experiments and Manipulation
- True experiment basics
- Involves controlling and changing the setup (manipulating the IV).
- Participants are typically assigned to different groups (experimental vs control).
- The primary aim is to establish causal relationships by isolating the effect of the IV on the DV.
- Example discussed: manipulating the amount of time spent on social media.
- Experimental group receives the treatment (e.g., time on social media).
- Control group does not receive the treatment.
- Non-experimental (observational) studies
- When researchers study things as they exist without random assignment or manipulation, they cannot definitively claim causality.
- These designs are often more efficient or ethical in certain contexts but limit causal inferences.
- Misinterpretation in the media
- Correlational findings are sometimes framed as causal in news reports, leading to overstatements.
- Researchers emphasize that correlation does not equal causation; additional evidence (including experimental data) is needed to assert causality.
- Experimental vs non-experimental data types
- Quantitative data: numbers collected from scales or ratings (e.g., happiness on a scale, frequency counts).
- Qualitative data: non-numeric information gathered from interviews or open-ended responses (to understand experiences and perspectives).
- Example related to “glasses” (experimental control concept)
- For a hypothetical test of an intervention, participants might be randomly assigned to wear glasses in the experimental group and not wear them in the control group to test effectiveness.
Data Types: Quantitative vs Qualitative
- Quantitative data
- Numerical data produced by scales or objective measurements.
- Examples from transcript: happiness rated on a scale, exercise frequency, time counts.
- Scales mentioned: a scale from 1 to 5; a scale from 1 to 10.
- Represented numerically; analysis often involves averages, variances, and statistical tests.
- Relevant ranges: 1 \,\leq x \,\leq 5 and 1 \,\leq x \,\leq 10 for the respective scales.
- Qualitative data
- Data derived from interviews, open-ended responses, and descriptive observations.
- Focus on how people think, feel, and experience phenomena rather than on numeric quantities.
- Analysis involves identifying themes across responses (thematic analysis).
Qualitative Research: Interviews and Thematic Analysis
- Qualitative methods often rely on interviews to explore attitudes, experiences, and perceptions.
- Example topic: attitudes toward aging and ageism in the workplace.
- Procedure described in transcript: conduct interviews with participants (e.g., “Have you ever experienced ageism in the workplace?”) and ask for details about experiences.
- Analysis approach: examine the interview data for recurring themes (e.g., prevalence of ageism among women, patterns across experiences).
- Purpose: understand how individuals experience phenomena and what those experiences reveal about broader social or psychological processes.
Examples, Implications, and Connections
- Example: Ice cream vs. crime correlation
- Demonstrates a correlation, not causation; potential third variables (e.g., heat, summer season) could influence both variables.
- Uses to illustrate the common misinterpretation of correlational findings as causal.
- Example: Time spent on social media and outcomes
- Demonstrates how researchers might manipulate an IV (time on social media) to observe effects on DV (e.g., mood, attention, etc.).
- Classroom observation scenario
- Observing sociability by counting how often children share in a preschool setting.
- The observational context (e.g., one-way mirror) reduces participant reactivity and aims to capture natural behavior.
- Research design connections to foundational principles
- Control of extraneous variables and random assignment in experiments is essential for causal inference.
- Observational methods provide ecological validity and descriptive insights but have limits for establishing causality.
- Ethical and practical considerations
- Minimizing observer effect and ensuring participants are not unduly influenced by being observed.
- Balancing the need for natural behavior with ethical responsibilities of researchers and educators in classrooms.
Terminology and Notation Recap
- Variables
- Independent Variable (IV): the treatment or condition manipulated by the researcher.
- Dependent Variable (DV): the outcome measured.
- Grouping and assignment
- Experimental group: receives the IV.
- Control group: does not receive the IV or receives a neutral condition.
- Random assignment: assigning participants to groups by chance to create equivalent groups at baseline.
- Data types
- Quantitative: numerical data from scales or counts.
- Qualitative: descriptive data from interviews or open-ended responses.
- Study designs
- Correlational study: examines relationships between variables without manipulation; cannot establish causality.
- True experimental study: manipulates IV and employs random assignment to infer causality.
- Conceptual takeaway
- Always distinguish correlation from causation; use experimental evidence to support causal claims; be mindful of potential third-variable confounds in observational data.
Quick Reference: Key Takeaways from the Transcript
- Correlation ≠ causation; third variables can influence observed relationships (e.g., genetics, environment).
- Observational methods (observer report vs naturalistic observation) are valuable for describing behavior but have limits in causal inference due to potential reactivity and lack of manipulation.
- True experiments require manipulation of IV and random assignment to establish causal effects; non-random designs provide correlational evidence only.
- Data can be quantitative (scales, frequencies) or qualitative (interviews, themes); both have distinct analytical approaches.
- Real-world examples cited in the transcript illustrate these concepts and common misinterpretations (e.g., media claims about causes based on correlational data).
Additional Notes (to supplement study prep)
- When designing a study in psychology, consider whether the research question requires causal conclusions or descriptive/relational insights.
- If ethical, practical, or logistical constraints prevent random assignment, rely on robust observational methods and triangulate with multiple data sources to strengthen inferences.
- In qualitative research, ensure systematic coding of themes and transparency about how themes were derived from interview data.
- When reporting findings, clearly articulate whether the study supports association or causation and discuss potential confounds and limitations.