KK

PSY 1001 – Chapter 2 (Research Methods Part 3) Comprehensive Notes

Experimental vs. Correlational Research

  • Experimental research uniquely identifies causal relationships.
    • Manipulate one variable (Independent Variable, IV) ➔ measure its effect on another (Dependent Variable, DV).
    • Central experimental question: Do people who experience different levels of the IV exhibit different behavior in terms of the DV?
  • Correlational research only measures two variables; no manipulation, therefore cannot establish causality.

Core Experimental Terminology

  • Independent Variable (IV)
    • The variable the researcher deliberately changes/manipulates.
    • Requires an explicit operational definition.
  • Dependent Variable (DV)
    • The outcome/behavior measured to assess IV’s influence.
    • Also must be operationally defined.
  • Experimental Group: Receives the IV treatment/condition.
  • Control Group: Does not receive the IV treatment (or receives a neutral/standard condition).
    • Note: Some experiments do not require a classic control group (e.g. comparing two active conditions).

Illustrative Experimental Scenarios

  • Drug X & Anxiety
    • IV: Ingestion of Drug X (Yes vs. No).
    • DV: Measured post-treatment anxiety level.
  • Music While Studying
    • IV: Study environment (Music vs. Silence).
    • DV: Test score on learned material.
  • Classroom Seating Arrangement
    • IV: Seating format (Rows vs. Circle).
    • DV: Number of students speaking during lecture.
  • Chemical Temperature Analogy
    • Demonstrates causal logic outside psychology: identical beakers heated vs. cooled leading to observable differences attributable solely to temperature.

Logical Steps to Infer Causation

  1. Correlation between IV & DV: DV changes when IV changes.
  2. Temporal Precedence: IV manipulation occurs before DV measurement.
  3. Elimination of Alternative Explanations: Remove or control all other variables that could influence DV.

Random Assignment vs. Random Sampling

  • Random Assignment
    • Procedure for distributing participants into experimental conditions by chance.
    • Creates statistically equivalent groups, compensating for individual differences.
  • Random Sampling
    • Procedure for choosing a subset from the target population.
    • Ensures sample represents population; separate from assignment within a study.

Designing the “Fancy Slides” Experiment (In-Class Example)

  • Research Question: Do students learn Intro Psych better with fancy colorful slides?
  • Procedure
    • Start with 100 students.
    • Randomly assign via coin-flip: Heads ➔ fancy slides (Liza); Tails ➔ plain slides (Mark).
    • Hold constant: instructor, lecture topic, classroom, time of day, test, etc.
  • Hypothetical Result: Fancy-slides group scores higher.
  • Potential Alternative Explanations (Confounds)
    • Different instructor/personality
    • Different lecture topic
    • Different room
    • Different time of day
  • Confound: Any factor that covaries perfectly with IV, offering another causal explanation.
    • Must be eliminated using experimental control (keeping all else equal).

Measurement Principles: Reliability & Validity

  • Reliability: Consistency of measurement across repeated trials.
    • Tape-measure example: multiple readings of 60 inches ➔ reliable.
  • Validity (Construct Validity): Degree to which measure captures the intended concept.
    • Tape-measure missing first 2 inches ➔ reliable but not valid (systematic error).
    • Extraversion survey containing happiness items ➔ unreliable indicator of extraversion’s true construct.
  • Key Insight: A measurement can be reliable without being valid; validity implies reliability, but not vice-versa.

Necessity of Statistics in Behavioral Research

  • Human behavior exhibits intrinsic variability; even identically treated individuals produce different outcomes.
  • Statistical tests help differentiate:
    1. True IV effect vs.
    2. Chance differences due to random variability.
  • p-value: Probability of obtaining observed group difference by chance if IV has no effect.
    • Standard threshold: p < .05 ➔ difference deemed “statistically significant.”

Ethical Principles in Research (Belmont-like Triad)

  • Beneficence: Maximize societal benefit, minimize participant harm (medical analogy: first, do no harm).
  • Autonomy (Respect for Persons): Secure informed consent; prohibit coercion.
  • Justice: Fair distribution of research burdens and benefits; avoid exploiting specific subpopulations.
  • Institutional Review Board (IRB) oversees adherence to these principles.

Psychologists as Scientists

  • Generate knowledge through empirical evidence.
  • Apply evidence-based treatments.
  • Throughout course topics, assess methods used to draw conclusions.

Recap & Practical Takeaways

  • Use random assignment and experimental control to isolate causal effects.
  • Vigilantly check for confounds; uphold reliability & validity in all measures.
  • Interpret research findings through appropriate statistical tests, primarily p-values.
  • Embed ethical guidelines throughout research lifecycle.
  • Remember: Psychologists apply a scientific lens; so should students evaluating psychological claims.