Name some non-experimental methods.
quasi-experiments, surveys, systematic observations, secondary data analysis, systematic reviews & meta-analysis
What do you do in correlational research?
measure and analyse the relationship between variables
Why may correlational research be used?
- not interested in cause and effect
- can’t manipulate predictor
Are variables (predictors/EVs) manipulated in correlational research?
no
How do quasi experimental designs differ from most experiments?
lacks full control: conditions are naturally-occurring & predictor can’t be isolated (to eliminate confounds)
Why may random allocation to conditions not always be possible in quasi-experiments?
- practicality/feasibility (e.g. can’t choose someone’s gender)
- ethics (e.g. can’t make people smoke to create a smokers condition)
What design should quasi-experiments try to use (if possible)?
pre-post designs
How do you carry out pre-post designs?
get a baseline measure & assess change before and after experimental manipulation
What are the strengths of quasi-experiments?
- good when random allocation isn’t possible
- can study real-world settings (schools, clinics, communities)
- ecological validity
- can show evidence of change over time
What are the limitations of quasi-experiments?
- selection bias - can systematically differ in unintentional ways
- attrition (dropout) - can systematically differ
- harder to separate cause and confound
- can’t isolate predictor (to eliminate confounds) SO can’t infer cause and effect
Name a type of quasi-experiment. (1)
natural experiment
What are the key features of natural experiments?
- no manipulation of predictor
- allocation to ‘conditions’ through external forces (existing traits/situations)
What are the strengths of natural experiments?
- high ecological validity
- can study rare/ethically-sensitive data
- often large-scale SO generalisable
- often have real-world relevance
What are the limitations of natural experiments?
- lack control over confounds
- rarely randomised SO low internal validity
- often low/lack replicability
- findings often context-specific SO limited generalisability/universality
- can’t isolate predictor (to eliminate confounds) SO can’t infer cause and effect
What is the advantage of surveys?
useful for large, diverse samples (time-efficient)
What are the disadvantages of surveys? (2)
- vulnerable to biases
- data quality is dependent on the question design and sampling strategy
Name the types of surveys. (2)
cross-sectional, longitudinal
cross-sectional survey
snapshot at one point in time
longitudinal survey
track the same sample over a period time
Name the survey modes. (3)
- online surveys
- face-to-face interviews
- telephone/postal surveys
What does the survey mode affect?
- who you can reach
- data quality
What are the advantages of online surveys?
cheap, fast, large reach
What is the advantage of face-to-face interviews?
rich detail
What are the disadvantages of face-to-face interviews?
expensive, labour-intensive, time-consuming
