Understanding Research Evidence
A practical overview of common research designs and what each can and cannot tell you.
Systematic reviews
Use predefined methods to identify and synthesize relevant studies. Quality depends on search strategy, inclusion criteria and the underlying evidence.
Randomized trials
Random allocation can reduce confounding and supports causal inference under appropriate design and execution.
Cohort studies
Follow groups over time to examine exposures and outcomes. Useful for many questions but vulnerable to confounding.
Case-control studies
Compare people with an outcome to those without it and examine prior exposures. Selection and recall bias can matter.
Cross-sectional studies
Measure exposures and outcomes at a point in time. Useful for prevalence but limited for establishing temporal sequence.
In-vitro research
Tests cells, molecules or biochemical systems. Valuable mechanistically, but results do not automatically translate to humans.