Experiments
- can infer causal (cause and effect) relationships
- has a control group and an experiment group
- expensive/costly, time-consuming, difficult to conduct
- Random assignment: it's a quasi-experiment without random assignment (e.g. two classes)
Correlational studies
- study the (negative/positive) relationship between two variables
- determines their correlation coefficient
- correlation doesn't suggest causation (illusory correlation)
Survey
- easy, not rigorous
- correlational design/data
- often not a random sample / non-representative sample (sampling bias)
Test
- Reliability: is it repeatable & yields similar results every time?
- Validity: does it measure what it wants to?
Naturalistic observations: exploratory
Case study: in-depth; for rare cases (not easy to do experimental manipulation)
Longitudinal study
- follows one group of people for a very long time
- very expensive, very difficult
- sampling bias: non-random withdrawl from the study; non-representative
Cross-sectional study
- study generational differences
- faster and easier compared to longitudinal design
- there might be a "cohort effect" (differences due to era, not aging/development)
Measures
- Performance (objective)
- Observational (e.g. interview)
- Self-report (e.g. questionnaires)