Notes: Research Methods & Statistics of Psychology (Chapter 2)

Theory, Hypothesis, and Research

  • Science is a methodology, not a discipline.
  • Essential elements: Theory, Hypothesis, Research.
  • THEORY: Explanation based on observations.
  • HYPOTHESIS: Prediction based on the theory.
  • RESEARCH: Test of the hypothesis; data are collected to evaluate the hypothesis.
  • Data either support the theory (leading to refinement) or refute the theory (leading to revision or discard).

Scientific Theory

  • A system of ideas that interrelates facts and concepts, summarizes existing data, and predicts future observations.
  • A good theory must be falsifiable.
  • Operational definitions specify the procedures used to produce or measure something.

Conceptual Level vs Concrete Level

  • Concepts (conceptual level): Hypothesized relationships among abstract ideas (e.g., Aggression, Frustration).
  • Concrete Level (operational definitions): Translate concepts into measurable variables (e.g., Number of times a child strikes a punching bag).

What Types of Studies Are Used in Psychological Research?

  • Three main designs: descriptive, correlational, experimental.
  • A variable is a quantity that can be measured or manipulated; quantification is central.

Descriptive Studies

  • Involve observing and describing behavior.
  • Naturalistic observation: passive observation.
  • Participant observation: researcher actively involved.
  • Developmental designs and observational records.
  • Advantages and limitations depend on the method; descriptive data provide rich descriptions but not causation.

Descriptive Studies: Naturalistic Observation

  • Observational records describe behavior in natural settings.
  • Advantages: rich descriptive data.
  • Disadvantages: limited to description; observer bias; anthropomorphic error (attributing human thoughts/feelings to animals).

Correlational Studies

  • Researchers do not manipulate variables.
  • Do not infer causation from correlations.
  • Strengths: ethical and practical for many questions.
  • Limitations: cannot establish cause-and-effect relationships.

The Directionality and Third Variable Problems

  • Directionality problem: A ↔ B correlation cannot specify which causes which (A → B or B → A).
  • Third variable problem: A may be related to B because of an unmeasured variable C.

Ethical Considerations for Correlational Designs

  • Some questions (e.g., war trauma) cannot be ethically studied by inducing conditions.
  • Correlational designs can address ethically permissible questions about associations.

Making Predictions from Correlational Research

  • Example: Depression is strongly related to suicide; correlation informs prediction but not causation.

Experimental Research Method

  • To identify cause-and-effect, experiments manipulate one or more independent variables and observe effects on a dependent variable.
  • Goal: determine what causes the observed behavioral differences.

Experimental Variables

  • INDEPENDENT VARIABLE (IV): Suspected cause.
  • DEPENDENT VARIABLE (DV): Outcome/measures of behavior.
  • Extraneous Variables: Other factors to be controlled to prevent confounding outcomes.

Groups and Selection

  • Experimental Group: experiences the IV manipulation.
  • Control Group: does not receive the IV manipulation.
  • Random sampling: Every person in the population has an equal chance of selection.
  • Random assignment: Each participant has an equal chance of being in either group.
  • Important for generalizability and reducing biases.

Population and Sampling

  • Population: the group you want to study (e.g., U.S. college students).
  • Sampling methods:
    • Random sample: taken at random from the population.
    • Convenience sample: taken from readily available subgroups.
  • Random assignment ensures comparability between groups.

The Psychology Experiment

  • Potential influences on results:
    • Research Participant Bias (e.g., placebo effects).
    • Placebo Effect: changes due to belief about treatment, not the treatment itself.
    • Placebo: fake treatment (e.g., sugar pill).
  • Experimental designs vary to offset biases.

Experimental Designs and Bias

  • Single-blind: participants unaware of hypotheses or group assignment.
  • Double-blind (not explicitly stated in slides, but implied as a fix): neither participants nor researchers know group assignments.

Data Collection Methods in Psychology

  • Determine the level of analysis: biological, individual, social, cultural.
  • Methods must fit questions at the chosen level of analysis.

Observational Data and Reactivity

  • Observational data: lab vs. natural environment.
  • Reactivity: participants altering behavior due to being observed (Hawthorne Effect).

Self-Report Methods

  • Ask people about themselves; interactive data collection.
  • Important to consider voluntary participation and honesty.
  • Example: drug-use surveys (voluntary).

Self-Report Bias

  • Social desirability: respondents answer to be viewed favorably.
  • Anonymity can reduce this bias.
  • Better-than-average effect: overestimating one's own abilities.

Response Performance Measures

  • Measure processing of information via task performance.
  • Major types:
    • Reaction time (speed of response).
    • Stimulus judgment (quality of judgment).
    • Response accuracy (correct vs. incorrect).
  • Advantages: simple, less observer bias.
  • Disadvantages: costly/time-consuming; may be less applicable to real-world settings.

Body/Brain Activity Measured Directly

  • Polygraph measures physiological indicators related to states.
  • Examples: heart rate, perspiration, blood pressure.

Psychophysiological Assessment: Brain Activity

  • EEG: measures electrical activity; produces electroencephalograms.
  • ERP: averages brain responses across trials to a stimulus.

Brain Imaging Modalities

  • PET: tracks metabolic activity via radioactive glucose.
  • MRI: strong magnetic field to visualize brain tissue.

Animal Research

  • Important data obtained from nonhuman animals.
  • Some research cannot be ethically conducted with humans; animals provide essential insights.

How Do We Know If We’re Wrong or Right?

  • Hypothesis testing requires falsifiability.
  • Statistics determine whether results are likely due to chance.
  • Findings with low probability under the null are deemed statistically significant.

Types of Statistics in Psychology

  • Descriptive Statistics: summarize data; describe distributions of scores.
  • Inferential Statistics: determine if observed differences reflect population differences beyond chance.

Descriptive Statistics: Summary of the Data

  • Two key ideas: Central Tendency and Variability.
  • Central Tendency measures: mean, median, mode.

Measures of Central Tendency

  • Mean: ar{x}
  • Median: middle value when data are ordered.
  • Mode: most frequently occurring value.

Mean and Its Sensitivity

  • The mean ar{x} is sensitive to extreme values; outliers can skew the distribution.

Measures of Variability

  • Range: difference between highest and lowest scores.
  • Standard Deviation: spread of scores around the mean; denote extSDextorσext{SD} ext{ or } \sigma.

Normal Distribution

  • Describes many natural phenomena; characterized by the mean and standard deviation.
  • Percentages around the mean follow a predictable pattern related to extSDext{SD}:
    • About 68% within ar{x} \, ext{±} \, ext{SD}
    • About 95% within ar{x} \, ext{±} \, 2\text{SD}
    • About 99.7% within ar{x} \, ext{±} \, 3\text{SD}

Inferential Statistics

  • Used to determine if observed differences between sample means reflect population differences or are due to chance.
  • Findings labeled as statistically significant when unlikely due to chance.

Correlation

  • Definition: a consistent, systematic relationship between two variables, measured by a correlation coefficient rr.
  • Possible directions:
    • Positive correlation: as one variable increases, the other increases.
    • Negative correlation: as one variable increases, the other decreases.
    • Zero/no correlation: no linear relationship.

Correlation Examples

  • Positive example: more activity often associates with more performance (conceptual example on slides).
  • Zero example: hair color and IQ (no reliable relationship).

Coefficient of Correlation (r)

  • Range: r ∈ [−1,1]r \, \in \, [-1, 1]
  • Sign indicates direction; magnitude indicates strength.
  • Perfect correlations are rare in psychology.

Interpreting Correlation Strength (rough guide)

  • Weak to moderate to strong relationships exist along a spectrum; exact thresholds depend on context.

Critical Thinking

  • Ask what evidence would support or refute the claim.
  • Gather relevant evidence relevant to the claim.

Four Basic Principles of Critical Thinking

  • Evidence quality varies; not all sources are equally reliable.
  • Authority or claimed expertise does not automatically make an idea true.

Astrology and Pseudopsychologies

  • Encourage testing claims; predictions often rely on uncritical acceptance.
  • Probability concepts can help evaluate predictive claims (e.g., base rates).