Psychological Inquiry and Research Methods
What this domain tests
This domain tests whether you can turn a psychological question into a defensible study. You need to operationalise variables, justify a design and sample, control extraneous variables, interpret data, and apply ethical and cultural safeguards. The exam rewards explicit links between method choices and the validity, reliability and credibility of conclusions - not just names of methods.
Core rules and concepts
- Operationalisation: define variables so they can be measured, e.g. sleep = minutes asleep measured by actigraphy, not “good sleep”.
- Designs: true experiments manipulate a variable and use random allocation; quasi-experiments lack random allocation; correlational designs measure co-variation without manipulation; case studies give depth but weak generalisation.
- Sampling: random, stratified, opportunity and snowball samples affect representativeness; volunteer bias is common.
- Validity and reliability: internal validity concerns confounds; external validity concerns generalisation; reliability concerns consistency. Demand characteristics and experimenter effects threaten both.
- Ethics: informed consent, voluntariness, avoidance of deception or full debriefing, confidentiality, protection from harm, and culturally safe practice such as consulting Māori communities and involving whānau where appropriate.
- Statistics: descriptive statistics summarise data; inferential tests estimate whether an effect is likely beyond chance. p < .05 is a convention, not proof; effect size and confidence intervals matter.
Worked example
A researcher asks whether sleep deprivation reduces memory. Aim: compare recall after 5 hours versus 8 hours of sleep. Operationalise sleep with actigraphy and memory with a 20-word free-recall task. Use an independent-measures design with random allocation, standardise testing time and caffeine intake, and counterbalance word lists. Reasoning: random allocation spreads individual differences; standardising caffeine reduces a confound; counterbalancing controls list difficulty. A significant difference supports a causal claim only if confounds are controlled. Ethical issues include informed consent, avoiding harm from sleep loss, and debriefing.
Traps
Correlation is not causation; a significant p-value is not a large effect; self-report is not behaviour; lab tasks may lack ecological validity; listing ethical principles without applying them to the specific study loses marks.