Please enable JavaScript.
Coggle requires JavaScript to display documents.
week 2 - causality and research designs (Cross sectional designs (little…
week 2 - causality and research designs
identifying the criteria of replicability, reliability and validity
Replicability
Transparent methods to allow for the theory to be tested more than once
Observations not bound to individual researcher
It could be done over
Reliability
Close to replicability, but has to do with outcomes of research
Tools for good study
Just to show that it's not a mere coincidence
Validity
Measurement validity; are the measurements used right?
Face validity
Convergent validity
Construct validity
Internal validity; are the (causal) relations found really there?
External validity; do findings hold for other cases than those studied?
Ecological validity; is the research setting similar enough to reality
Causality
Criteria
Plausibility (good reason why cause may lead to effect)
Temporality (cause comes before effect)
Consistency (this always happens)
Causation A->B
Reverse causation A<-B
Intervending effect A->C->B
Cyclic effect A->B->A->B etc
Confounding effect C->(AB)
Spurious relation A&B (mere coincidence)
Research designs
Experimental designs
The quintessential model for (quantitative scientific research)
Comparative nature
Two or more groups -> 1 control and 1 manipulated
1) Laboratory experiment
Excluding any confounding effects (high internal validity)
'sterile' (low ecological validity)
peole are aware of being part (reactive effects)
2)Field experiment
More realistic (high ecological validity)
Less control over confounds (high internal validity)
Cross sectional designs
little manipulation
Take a cross-section of reality
Can't really be sure about causal relations
So - internal validity low, external validity high
Longitudinal designs
Cohort = different people, same criterion
panel = single group of people studied repeatedly