Study type | Potential goals | Advantages | Disadvantages |
---|---|---|---|
Cross-sectional | - Population prevalence of exposure and/or outcome | - Relatively simple | - Poor for rarer exposures and outcomes |
- Associations between exposures and outcomes | - Relatively cheap | - No causality may be inferred as exposures and outcomes are measured contemporaneously | |
- Relatively quick | - Highly susceptible to information bias | ||
- Good for common conditions and exposures | |||
- May assess multiple exposures and outcomes | |||
- Good for initial assessment of an exposure or outcome | |||
Case–control | - Associations between exposures and outcome | - Relatively cheap | - Choice of controls notoriously difficult |
- Strength of association in the form of odds ratio between exposure(s) in controls and exposure(s) in cases | - Relatively quick | - May only examine one outcome | |
- May assess long latent periods | - Odds ratio not an intuitive measure | ||
- Good for rarer outcomes | - Highly susceptible to selection and information bias and population stratification | ||
Cohort | - Incidence rates | - Good for rare exposures | - Not simple |
- Temporal associations between exposures and outcomes | - May examine multiple exposures and outcomes | - Not cheap | |
- May assess long latent periods | - Not quick (unless retrospective) | ||
- May assess temporal relationship between exposure and outcome inferring causality | - Highly susceptible to retention bias | ||
- Susceptible to sampling bias |