Psychology: Research Methodology Notes

Experimental Methodology

  • Involves use of an experiment with independent variable(s) and random assignment to groups.

The Scientific Method in Psychology

  • System used to reduce bias and error in measurement of data.
  • Core steps: Perceive → Hypothesize → Test → Draw conclusions → Report, revise, replicate.
  • Typical sequence: Perceive and Hypothesize (detect a phenomenon and propose a testable explanation) → Test (conduct experiment) → Report (share results) → Revise/Replicate (evaluate validity and repeat).
  • Example flow: Observe brother plays and becomes more aggressive after watching violent cartoons on Saturday; Hypothesize that violent cartoons increase aggression; design an experimental scenario; test results; determine if hypothesis is supported, if test is valid, and if confounding variables were controlled; report to teachers/school/community; consider revisions and replication.

Perceive and Hypothesize

  • Observe a phenomenon and propose a testable explanation.
  • Example hypothesis:
    • If kids are exposed to violent cartoons on a regular basis then they will become more aggressive in their everyday lives.
  • Considerations when testing: validity of the test, control of confounding variables, replication potential, and reporting results to stakeholders.

Sampling and Representativeness

  • Representative Sample From Population to Sample: a random sample of subjects from a larger population; each member has an equal chance of being selected.
  • Sampling bias: occurs when some members of a population are systematically more likely to be selected.
  • Convenience sampling: a type of sampling bias where units are chosen for ease of access (e.g., counselor needing students at 4:00 pm for a survey).
  • Other biases:
    • Geographical proximity
    • Willingness to participate
    • Availability at a given time
  • If the sample is not random, problems arise in generalizing results.

Hypothesis and Falsifiability

  • Hypothesis must be falsifiable: it should be possible to conceive an observation that could disprove it.
  • Example falsifiable claim: “Aliens do not exist.” (Could attempt to find evidence to prove wrong.)
  • A good experimental hypothesis: If kids are exposed to violent cartoons on a regular basis then they will become more aggressive in their everyday lives.

Experimental Methodology: Key Terms

  • Independent Variable (IV): the variable that is deliberately changed.
  • Dependent Variable (DV): the variable that is measured.
  • Operational Definition: precise, measurable definitions of variables to allow replication.
  • Confounding Variables: variables that interfere with the interpretation of the IV's effect.
  • Random Assignment: assigning participants randomly to either the control or experimental group.
  • Control Group: does not receive the IV.
  • Experimental Group: receives the IV.

Operational Definitions and Measurement

  • Operationally define levels of a variable to allow replication (e.g., level of fear in a dog experiment): low, moderate, high.
  • Clear operational definitions improve replicability.
  • Definition example: aggression on a playground could be coded as: hands on another kid, pushing, slapping, hitting, etc.; but what about a yell? Define each behavior precisely.

Confounding Variables and Control

  • Confounding variables can affect outcomes; researchers strive to control for them or randomize to distribute them evenly across groups.
  • Methods to control for confounds include random assignment, blinding, and standardized procedures.

Random Assignment, Blinding, and Biases

  • Random Assignment: random allocation of participants to control or experimental groups.
  • Single Blind: participants do not know if they are in control or experimental group.
  • Double Blind: neither participants nor researchers know who is in which group.
  • Experimenter Effect: researcher unintentionally influences participants or data to arrive at a preferred outcome.
  • Placebo Effect: participants’ beliefs about treatment influence outcomes when the treatment has no therapeutic effect.
  • Social Desirability Bias: respondents answer in ways that are socially acceptable rather than truthful.
  • Strategies to control confounds: use single/double blind designs, rigorous protocols, standardized interactions, and objective measures.

Experimental Terms and Definitions

  • Random Assignment: random selection of who will be in the experimental vs. control group.
  • Single Blind: participants do not know group assignment.
  • Double Blind: neither participants nor experimenters know group assignment.
  • Experimenter Effect: bias in data collection or interpretation toward expected results.
  • Placebo Effect: belief in benefit leads to perceived improvement even without active treatment.
  • Social Desirability Bias: responses align with societal expectations rather than true beliefs.

Placebo and Video Example

  • Placebo Effect in Videos: Blooket Game 0.2 - a patient took placebos for over five years; illustrates how beliefs can influence outcomes even without active treatment.

Measuring Variables: Qualitative vs. Quantitative

  • Qualitative: descriptive, not measured numerically (e.g., color, sweetness, perceived quality).
  • Quantitative: measured with numeric values (e.g., Likert scales, test scores).

Validity and Reliability

  • Validity: the degree to which an experiment measures what it is supposed to measure.
  • Reliability: the tendency of an experiment to yield the same results under the same conditions.
  • Peer review and replication are important for confirming conclusions.
  • IQ Example:
    • IQ is reliable and valid.
    • IQ is reliable but invalid.
    • IQ is unreliable and invalid.
    • (Analogy: using a dartboard: repeatedly hitting the same spot does not guarantee it’s the true target.)

Non-Experimental Methodologies

  • Naturalistic Observation: watching animals or people in their natural environments; yields realistic behavior but lacks control.
    • Potential issues: observer effect, participant observation, observer bias.
  • Laboratory Observation: observing in a controlled, artificial setting; allows more control but may reduce natural behavior.
  • Case Studies: in-depth investigations of one subject; findings may not generalize to others (e.g., Phineas Gage, DID).
  • Surveys: standardized questions to large groups; issues include self-report bias and wording/order effects.

Correlation and Meta-Analysis (Non-Experimental)

  • Correlation: measures the statistical relationship between two variables with little to no control of extraneous variables.
  • Meta-Analysis: quantitative synthesis of results from multiple studies.
  • Potential advantages: improved precision, answers to more questions, helps resolve controversies.

Measuring Relationships: Correlation Details

  • Finding Relationships: Correlation involves two variables (X and Y).
  • Example: Variable 1: Smoking, Income, Education; Variable 2: Health.
  • Correlation Coefficient r: represents direction and strength of the relationship.
  • Positive correlation: r > 0; Example: a relationship where increases in one variable accompany increases in the other.
  • Negative correlation: r < 0; Example: one variable increases as the other decreases.
  • Critical reminder: Correlation does not imply causation.
  • Spurious correlations example ( Ice Cream and various outcomes ): illustrate non-causal associations that can arise from third variables like season or population behaviors.
  • Mathematical representation:
    • \begin{equation} r = \frac{\text{cov}(X,Y)}{\sigmaX \sigmaY} \end{equation}
    • Positive correlation: r > 0; Negative correlation: r < 0; No correlation: r \approx 0.

Ethics in Psychological Research

  • Protection of rights and well-being of participants.
  • Informed consent and the right to withdraw at any time.
  • Justification for deception when used; debriefing after the study.
  • Confidentiality and data protection.
  • Debriefing and correcting any undesirable consequences.
  • Peer review and ethical oversight to minimize harm.

Ethics in Animal Research

  • Animal research addresses questions not easily answerable in humans.
  • Focus is on avoiding unnecessary pain and suffering.
  • Animals are used in a minority of studies (approximately 7%).

Additional Resources and References

  • Note: Some slides reference external resources such as Top Unethical Experiments lists and AP/ED materials for newer curricula; these provide context to historical and modern ethical standards.

Quick Reference: Core Concepts to Remember

  • IV and DV definitions; operational definitions for replication.
  • Random assignment, control vs. experimental groups.
  • Single vs. double blindness; placebo and experimenter effects.
  • Sampling methods: representative random samples vs. convenience samples.
  • Qualitative vs. quantitative data.
  • Reliability and validity and their importance for credible conclusions.
  • Non-experimental methods: naturalistic observation, case studies, surveys, correlation, meta-analysis.
  • Correlation does not imply causation; use caution with spurious relationships.
  • Ethical principles: informed consent, deception, withdrawal, confidentiality, debriefing.
  • Animal research ethics and alternatives where possible.

End of Notes