Research Methods and the Scientific Method in Psychology

Foundations of Scientific Methodology in Psychology

  • Definition of Science: A discipline is classified as a science based on whether it utilizes the multistep scientific method to guide a researcher's investigations.
  • Purpose of the Scientific Method: It is the only impartial and objective method for acquiring factual knowledge and truth, free from bias. Early psychological pioneers were directly influenced by philosophers who championed this structured approach to establish empirical facts.
  • Key Characteristics of the Scientific Method:
    • Deductive Nature:
    • Science operates deductively, not inductively.
    • Induction begins at the bottom by observing data patterns and attempting to construct theories to explain them.
    • Flaw of Induction: Starting with data yields numerous competing explanations or theories that all account for the exact same pattern of data, leaving researchers with no clear way to determine which explanation is correct.
    • Florida State University (FSU) 1970s Study Example: A very attractive male research assistant approached female coeds on campus asking for sex later that evening, and a very attractive female research assistant did the same with male coeds. Sex did not actually take place; researchers tested compliance rates.
      • Results: Approximately 90%90\% of male coeds agreed to sex with a stranger, whereas only about 5%5\% of female coeds agreed.
      • Inductive Exercise Outcome: Brainstorming explanations for this data yields 7, 8, or 9 competing theories (e.g., social norms, evolutionary drives, safety concerns) prior to evaluating the researchers' specific evolutionary theory.
    • How Deduction Operates: A researcher begins with a single theory (with no competing theories at the start). Testable hypotheses are derived from this theory and subjected to controlled laboratory experiments.
    • Falsification Principle: If laboratory or real-world results fail to support the hypothesis, the underlying theory must be abandoned or revised immediately.

Observational Approaches for Theory Formulation

  • Role of Observational Studies: Descriptive and observational approaches are utilized at the standard beginning of the scientific method to generate initial theories.
  • Case Study Method:
    • Involves an in-depth, structured study of a single individual.
    • Focuses on atypical individuals exhibiting rare behaviors or unique experiences rather than an average person.
    • Case Example: A 14-year-old girl who was chained up and isolated for 14 years, deprived of all sensory inputs (seeing and hearing).
    • Significance: Provided a rare, non-experimental opportunity to evaluate whether critical periods or deadlines exist for acquiring human skills like language and emotional attachment, without ethically subjecting a child to intentional deprivation.
  • Hypothetical Seating Row Case Study:
    • Investigation of student performance origins: A student who earned 60s/D's across middle school shifted to earning straight A's throughout high school.
    • In-depth structured interviews revealed that this grade shift coincided directly with a physical relocation from the back row to the front row on the first day of 9th grade, where he remained seated for 40 years.
    • Theory Derived: The physical row or distance a student sits relative to the teacher directly causes student performance.
  • Naturalistic Observation Method:
    • Behavioral observation conducted in natural settings without researcher intervention or subject awareness.
    • Classroom Observation Study: Two independent raters observe a live classroom to measure specific performance indicators:
    • Number of hand raises.
    • Total words typed on laptops.
    • Minutes spent gazing at the whiteboard.
    • Pop-quiz performance scores.
    • Outcome: Raters record zero performance behaviors among back-row students, while front-row students display high frequencies of these behaviors, generating theories regarding physical location and academic output.

Correlational Methodology and Predictive Analysis

  • Testing Theories: Theories are scientifically valuable only if they generate testable predictions. Untestable assertions remain speculation or conjecture.
  • Hypothesis: A testable prediction generated directly by a theory.
  • Correlational Approach:
    • Conducted specifically to form hypotheses and make accurate predictions.
    • Correlation: Refers to a statistical relationship between two variables.
    • Two distinct correlation types exist (Textbook Section 2.3):
    • Positive Correlation (Direct Relationship):
      • As one variable increases in value, the other variable increases correspondingly.
      • Does not connote a "good" outcome.
      • Graphical Representation: Data points slope upward from left to right.
      • Height and Weight Example: Plotted across 100 lbs100\,\text{lbs} to 400 lbs400\,\text{lbs} and 5 ft5\,\text{ft} to 7 ft7\,\text{ft}. Higher weights directly track with taller statures, establishing a predictable direct relationship without asserting a direct causal mechanism.
      • SAT Scores and College GPA Example: Colleges mandate SAT/ACT scores because SAT measures aptitude (the predictive ability to learn), not prior achievement. SAT scores predict later college GPA, which determines degree completion over 4 years (8 semesters).
      • An SAT score of 800800 predicts a GPA under 2.02.0 (ineligibility to graduate, leading to dropout within ~3 semesters).
      • An SAT score of 12001200 predicts a GPA sufficient for graduation.
      • Study Duration and GPA Example: Plotting study time (0.5 hours/week0.5\,\text{hours/week} vs. 14–15 hours/week14\text{--}15\,\text{hours/week}) against GPA shows that minimal study time predicts a GPA of  ≈1.0 \,\approx 1.0\, (D grade), whereas studying 14–15 hours/week14\text{--}15\,\text{hours/week} predicts a GPA of  ≈3.0 \,\approx 3.0\, or higher.
    • Negative Correlation (Inverse Relationship):
      • As one variable increases, the other variable decreases correspondingly.
      • Does not connote a "bad" outcome; predictive power is equal to that of a positive correlation.
      • Age and Fun Example: Early childhood (low rules and responsibilities) correlates with high levels of fun; as age increases along with financial obligations and parental duties, fun decreases.
    • Zero / No Correlation:
      • Occurs when two variables lack any directional relationship.
      • Graphical Representation: Appears as a random distribution of data points across a scatter plot.
      • Steelers Wins vs. Student Grades Example: Plotting seasonal wins by the Pittsburgh Steelers (ranging from 11 to 1717 wins) against the number of students receiving A's in a psychology class yields a random distribution. Winning 88 games corresponds variously to low, moderate, and high student A counts, offering zero predictive capability.

Survey Design, Response Biases, and Sampling

  • Hypotheses in Seating Row Studies:
    • Null Hypothesis: Seating position has no effect; front-row performance equals back-row performance (Front=Back\text{Front} = \text{Back}) on a 70-question final exam.
    • Researcher Hypothesis: Front-row seating causes superior performance; front-row scores exceed back-row scores (Front>Back\text{Front} > \text{Back}).
  • Survey Methodology in Correlational Research (Textbook Section 2.2):
    • Involves administering questionnaires that collect historical or self-reported data (e.g., asking past classroom row location and past grades earned).
    • Social Desirability Bias: Subjects systematically alter answers to present themselves favorably in research settings.
    • Example: A pro-choice respondent answering a pastor-administered survey may mark pro-life choices due to fear of social judgment despite guaranteed confidentiality.
    • Response Tendencies:
    • Yea-saying: The tendency of survey respondents to answer "yes" or select maximum positive values (e.g., rating everything a 77 on a 1–71\text{--}7 Likert scale) indiscriminately across all questions.
    • Nay-saying: The tendency to answer "no" or select minimum values (e.g., rating everything a 11) indiscriminately.
    • Methodological Safeguards Against Survey Bias:
    • Catch / Lie Items: Inserting statements designed to detect dishonest self-presentation, such as marking "true" to "I've never lied in my life, never even a white lie." Answering "true" invalidates the respondent's data.
    • Rephrased Duplicate Items: Asking the exact same question twice in different parts of the survey with distinct wording. Inconsistent answers ("yes" to one, "no" to the other) identify inattentive respondents.
  • Population Definition:
    • Target Population: The complete set of individuals to whom a specific research question applies.
    • For seating row research, the population comprises every person worldwide who has ever taken a face-to-face class (≈3–4×109\approx 3\text{--}4 \times 10^9 individuals out of a total global population of 8×1098 \times 10^9).

Causal Limitations and the Third Variable Problem

  • Core Principle: Correlation does not equal causation. Even a perfect correlation (r=±1.0r = \pm 1.0) between two variables cannot prove that one variable causes the other.
  • The Third Variable Problem: Observed changes in two correlated variables may be driven entirely by an unmeasured, confounding third variable.
  • Confounding Variables in Classroom Seating Studies:
    • Sensory Ability / Vision (Textbook Chapter 5): Physical laws of optics dictate that light waves reflected from a front screen project a significantly larger image on the retinas of front-row students compared to back-row students.
    • Teacher Attention: Instructors inherently pay greater attention, check notebooks, and provide direct positive reinforcement to front-row students while ignoring or criticizing back-row students.
    • Distractions: Back-row seating features higher concentrations of peer chatter, off-topic behavior, and social distractors.
    • Student Motivation and Social Stress: Intrinsic differences in baseline academic drive and stress levels systematically differ between students who self-select into front versus back rows.

Experimental Design and Causal Inference

  • Experimental Superiority in Establishing Causation:
    • Unlike correlational designs that evaluate past behavior via retrospective self-reports, experiments test participants actively in real-time under strictly controlled conditions.
    • Random Assignment:
    • Randomly placing participants into experimental conditions equates the groups prior to testing on every possible participant dimension (e.g., intrinsic motivation, baseline intelligence, sensory visual acuity, social stress levels).
    • Methodological Isolation:
    • Experimental protocols ensure both groups receive identical treatment (equal instructor attention, identical distraction levels, equal praise) except for one single factor.
  • Experimental Variables:
    • Independent Variable (IV): The single factor manipulated directly by the experimenter to test its specific impact. It stands alone as the isolated difference between groups.
    • Dependent Variable (DV): The variable measured by the experimenter to assess the outcome or effect produced by the independent variable.
  • Milgram Obedience Experiment Example (Textbook Section 12.4):
    • Stanley Milgram deceived adult male participants into believing they were delivering lethal electric shocks to an unseen stranger to test obedience toward an authority figure wearing a lab coat.
    • Demonstrates strict manipulation of independent variables alongside significant historical ethical considerations.
  • Experimental Analysis Framework:
    • Evaluating experimental descriptions requires identifying four critical parameters:
    1. The Independent Variable (IV).
    2. The Dependent Variable (DV).
    3. The controlled or equated variables across groups.
    4. The confounded variables that threaten internal validity.