Research Methodology Notes - Ch. 1
The Research Process
- Psychologist's Research Steps: Aim/hypothesis, method selection, variable definition/control, ethics, participants, data analysis/conclusions, evaluation.
- Why Research? To test hypotheses and find evidence.
Research Methods Overview
- Three Main Types:
- Experimental: Lab, Field, Quasi/Natural. Can show cause-and-effect.
- Non-experimental: Self-reports, observations, case studies, correlations. Cannot show cause-and-effect. Can be used alone or with experimental for extra info.
- Longitudinal: Can be experimental or non-experimental; lasts longer, tracks changes over time.
EXPERIMENTAL RESEARCH METHODS
- Goal: Determine if an Independent Variable (IV) causes changes in a Dependent Variable (DV).
- Method: Create 2+ conditions (e.g., low vs. high background light) and measure effects on behavior.
- Causality: Requires controlling extraneous variables.
- Experimental Design: Compares experimental conditions (groups receiving IV) to a control condition (no IV).
- Experimental groups: Receive IV.
- Control group: Comparison, no IV.
1. TYPES OF EXPERIMENTAL RESEARCH METHODS
All aim to show cause-and-effect.
A. Laboratory Experiments
- IV impacts DV. Participants assigned to design (independent or repeated measures).
- Setting: Artificial, highly controlled lab.
- Pros: Fewer extraneous variables, high standardization, easy replication.
- Cons: Lower ecological validity/mundane realism, researcher bias risk.
B. Field Experiments
- Setting: Participant’s natural environment.
- Pros: Higher ecological validity, lower demand characteristics, natural behavior.
- Cons: More uncontrolled variables, harder to replicate/check reliability, ethical issues (consent, deception).
C. Quasi/Natural Experiments
- Uses naturally occurring IVs (e.g., gender, age).
- No random assignment; grouping is natural.
- Causality: Harder to establish.
- Strengths: Real-world relevance.
- Weaknesses: No random assignment, participant variables, hard to infer causality.
TYPES OF NON-EXPERIMENTAL RESEARCH METHODS
Cannot show causality, but describe phenomena or reveal relationships.
A. Self-Reports (Questionnaires and Interviews)
- Questionnaires: Written. Open/closed questions (Likert, rating scales).
- Pros: Quick, efficient, potentially truthful.
- Cons: Social desirability, memory issues.
- Interviews: Spoken. Structured, semi-structured, unstructured.
- Pros: Reveal reasons, in-depth insight.
- Cons: Interviewer bias, social desirability, memory issues.
B. Observations
- Directly observing behavior. Overt/covert, participant/non-participant, structured/unstructured, naturalistic/controlled.
- Measurements: Time sampling (fixed intervals) or event sampling (every occurrence).
- Strengths: Captures actual behavior, high ecological validity (natural).
- Weaknesses: Observer bias, subjectivity, time-consuming.
C. Case Studies
- In-depth analysis of individuals/small groups using various methods (interviews, questionnaires, observations).
- Usually longitudinal; no variable manipulation.
- Strengths: Rich, detailed data; ecological validity.
- Weaknesses: Hard to generalize, observer bias, low reliability, hard to replicate.
D. Correlations
- Explore relationships between variables, not causation.
- Types: Positive (direct), negative (inverse), zero/no.
- Strengths: Identify relationships for further study, fewer ethical concerns.
- Weaknesses: Cannot infer causality, mainly quantitative data (unless combined). Correlation e causation.
LONGITUDINAL RESEARCH
- Follows one group over time, assessing variables at intervals.
- Can be experimental or non-experimental.
- Advantages: Tracks individual changes over time (self-comparison), controls for situational variables.
- Disadvantages: Generalizability issues due to attrition (dropouts), researchers/measurements may change (outdated), ongoing consent needed.
EXPERIMENTAL DESIGN (RESEARCH DESIGN)
- Only for experiments; how subjects are allocated to IV levels/groups.
A. Independent measures (between-subjects)
- Different subjects for each IV level; compare data between groups.
- Random assignment or matching to control participant variables.
- Strengths: Lower demand characteristics, fewer carry-over effects.
- Weaknesses: Participant variables, more participants needed.
- Matching: (Matched Pairs) reduces participant variables.
B. Repeated measures (within-subjects) - Same participants in all IV levels; compare data within individuals.
- Strengths: Each person is own baseline, reduces participant variables, fewer participants needed.
- Weaknesses: Order effects (practice, fatigue), demand characteristics.
- Control: Counterbalancing and randomization for order effects.
RESEARCH METHODS COMPARISON
- Experimental: Show cause-and-effect (Lab, Field, Quasi/Natural). Need an experimental design.
- Non-experimental: No causality (Self-reports, Observations, Case Studies, Correlations). No fixed research design.
VARIABLES AND OPERATIONALIZATION
- DV (Dependent Variable): What is measured.
- IV (Independent Variable): What the researcher manipulates to create groups/conditions.
- Ask: DV = what is measured? IV = how are groups different?
- Operationalizing: Turning abstract concepts (e.g., helpfulness) into measurable indicators.
- DV on Y-axis for graphs.
STANDARDIZATION, CONTROLS, AND EXTRANEOUS VARIABLES
- Why control? Ensure only IV affects DV; strengthens causal inference.
- Standardizing procedures: All participants follow same sequence, instructions, materials, location.
- Benefits: Easier replication, reliability, increases validity.
- Controls vs. control conditions:
- Controls: Procedures minimizing extraneous variables.
- Control conditions: Baseline for comparison.
- Extraneous vs. confounding variables:
- Extraneous: Other variables influencing DV (threats to validity).
- Confounding: Extraneous variables with a systematic effect, distorting IV-DV relationship.
- Pilot studies: Small pre-studies to identify variables needing control.
- Types of extraneous variables:
1) Situational: Environment (temperature, noise).
2) Participant: Traits (intelligence, mood, sleep).
3) Order effects: (In repeated measures) practice, boredom, fatigue.
4) Experimenter effect: Bias, expectations of researcher.
5) Demand characteristics: Participants guess study aim, alter behavior. - Controlling extraneous variables: Counterbalancing, random allocation, matching, standardized procedures, double-blind designs.
- Double-blind tests: Neither participant nor researcher knows condition.
AIMS AND HYPOTHESES
- Aim: Study purpose, stated before data collection.
- Hypotheses: Testable, falsifiable, concise statements with IV and DV. "If (IV manipulation) then (DV outcome)."
- A. Alternative (H1) hypotheses:
- Directional (one-tailed): Predicts specific direction (e.g., increase/decrease).
- Non-directional (two-tailed): Predicts a difference, but not direction.
- B. Null hypothesis (H0): Predicts no effect/difference; any difference is due to chance. Aim is to reject H0. Example: "There will be no difference between groups."
- Theory vs. Hypothesis:
- Hypothesis: Testable prediction for specific study.
- Theory: Well-substantiated explanation for verified facts, supported by evidence.
SAMPLING METHODS
Sample: Smaller group representing wider population.
Population/Target Population: Larger group of interest.
Purpose: Recruit representative sample for generalizability.
A. Opportunity sampling: Use readily available people.
- Strengths: Quick, large numbers.
- Weaknesses: May not be representative.
B. Volunteer (self-selecting) sampling: Participants volunteer via ads. - Strengths: Lower drop-out rates.
- Weaknesses: Not representative (self-selection bias).
C. Random sampling: Every population member has equal chance of selection (gold standard). - Strengths: Higher representativeness.
- Weaknesses: Difficult to obtain participation/reach individuals.
TYPES OF DATA
- Data: Results collected.
- Quantitative data (QT): Numerical measurements.
- Strengths: Comparison, statistics, objective.
- Weaknesses: Oversimplifies behavior, limited to numbers.
- Qualitative data (QL): Descriptive, word-based.
- Strengths: Captures explanations, meanings.
- Weaknesses: Subjective interpretation, researcher bias.
- Quantitative data (QT): Numerical measurements.
DATA ANALYSIS – STATISTICS
- Normal distribution: Symmetrical, bell-shaped; mean, median, mode are equal.
- Skewed distribution: Asymmetrical data.
- Measures of central tendency:
- Mean: ar{x} =\frac{1}{n} extstyle \int\sum{i=1}^{n} xi
- Median: Middle score.
- Mode: Most frequent score.
- Measures of spread:
- Range: \text{Range} = x{\text{max}} - x{\text{min}}
- Standard deviation (SD):
- Smaller SD = scores closer to mean.
- Larger SD = more variability.
- Graphs:
- Bar chart: Categorical data, separated bars.
- Histogram: Continuous data, bars touch.
- Scatter plot: Plots IVs vs. DVs; shows relationships, trend line, outliers.
- Strong positive correlation: Points close to line of best fit.
- Moderate/weak: More scatter.
- Validity/Reliability (overview):
- Validity: Measures what it intends to, accurate conclusions.
- Reliability: Consistency of results.
- Transferability/Generalizability: Apply findings to other contexts.
- p-value: (AICE note) "p value lower than 0.5" for rejecting null hypothesis (conventional is 0.05). Not typically asked on AICE.
RELIABILITY AND VALIDITY
- Reliability: Consistency of study/measurement. Replicable? Same results?
- Improve: Standardize procedures, control extraneous variables, consistent data collection.
- Assess: Test-retest (same results twice), Inter-rater (observers agree).
- Qualitative methods have limited replication reliability.
- Validity: Measures what it's supposed to, generalizable results.
- Control: How well extraneous variables are controlled.
- Realism (ecological validity): Natural behavior? Minimized demand characteristics? Realistic task?
- Generalizability: Applicable beyond sample/setting?
ETHICS OVERVIEW
- Why? Respect participants' rights and dignity (BPS, 2014).
- Core categories: Humans, Animals.
- Key principles: Respect, informed consent, right to withdraw, protection from harm, confidentiality, privacy, debriefing.
ETHICS GUIDELINES RELATED TO HUMANS
- Informed consent: Sufficient, clear info for decision.
- Right to withdraw: Participants can leave anytime, data removed; no coercive incentives.
- Protection: No physical/psychological harm; end in similar state as started.
- Deception: Only if necessary/justifiable, no distress likely.
- Confidentiality: Personal data private; use anonymity.
- Privacy: Right to ignore questions; observe only expected behavior.
- Debriefing: Post-study explanation; mitigate negative mood.
ETHICS GUIDELINES RELATED TO ANIMALS
- Replacement: Use simulations if possible.
- Species/strain: Choose least suffering risk.
- Number: Smallest necessary.
- Pain/distress: Avoid physical/psychological pain.
- Housing: Respect species needs.
- Reward/deprivation/aversive stimuli: Prefer rewards; avoid deprivation/aversive.
- Anesthesia, analgesia, euthanasia: Use anesthesia/analgesia; humane euthanasia if needed.
EVALUATING RESEARCH – RELIABILITY AND VALIDITY (RECAP)
- Reliability: Replicable with same methods/data? Standardization helps.
- Validity: Measures what it intends? Conclusions warranted? Focus on internal, external, construct validity.
RESEARCH METHODS GRAPHIC ORGANIZER
- Visual tool mapping Experimental vs. Non-Experimental methods and longitudinal applicability.
ADDITIONAL NOTES
- Methodology question (exams): Evaluate researcher decisions (design, sampling, validity/reliability, generalizability, ethics) and impact on results.
- General Study Path: Understand question, select method, specify IV/DV, controls/standardization, ethics, sampling, data types, analysis, anticipate validity/reliability.
- Notation Recap:
- IV: Manipulated variable.
- DV: Measured outcome.
- Extraneous: Unwanted variables affecting DV.
- Confounding: Systematic extraneous bias.
KEY FORMULAS AND IDENTIFIERS
- Mean: \bar{x} =\frac{1}{n}
- Range: \text{Range} = x{\text{max}} - x{\text{min}}
- Standard deviation (sample): s (Use standard formula with sum of squared deviations from mean).
- Hypotheses structure: "If (IV manipulation) then (DV outcome)."
- Null hypothesis: "No difference between groups; any difference due to chance."
- p-value: