Types of study design
Reading File
Finding Sources
Searching PubMed
"Research Design"[MeSH Terms]
Searching the Web
types of epidemiological study designs CDC observational experimental
Types of study design
Research studies are broadly classified into observational and experimental (interventional) studies. Reviews of existing research are usually considered separately.
1. Observational studies
The researcher does not assign an exposure or treatment. They observe what occurs naturally.
| Design | What it does | Typical use |
|---|---|---|
| Case report | Detailed description of one unusual patient/event | Identifies a new or rare condition, adverse effect, or presentation |
| Case series | Describes several similar cases, without a comparison group | Generates hypotheses |
| Cross-sectional study | Measures exposure and outcome at one point in time | Estimates prevalence, such as prevalence of hypertension in a community |
| Case-control study | Begins with people who have the outcome (cases) and compares prior exposures with controls without it | Useful for rare diseases or diseases with long latency; usually produces an odds ratio |
| Cohort study | Begins with exposed and unexposed groups and follows them for outcome occurrence | Measures incidence and relative risk; can be prospective or retrospective |
| Ecological study | Uses data for groups or populations rather than individuals | Useful for population-level hypotheses but prone to ecological fallacy |
Cohort study subtypes
- Prospective cohort: exposure is measured now and participants are followed into the future.
- Retrospective cohort: existing records are used to reconstruct exposure and follow-up.
- Ambispective cohort: includes both retrospective and prospective components.
2. Experimental or interventional studies
The researcher assigns an intervention, such as a drug, vaccine, or health education program.
| Design | Main feature |
|---|---|
| Randomized controlled trial (RCT) | Participants are randomly allocated to intervention and control groups. Best design for assessing treatment efficacy and causal effects. |
| Non-randomized controlled trial / quasi-experimental study | Intervention is assigned but randomization is absent. |
| Before-and-after study | Outcomes are measured before and after an intervention in the same group. |
| Crossover trial | Each participant receives more than one treatment in sequence, usually with a washout period. |
| Cluster randomized trial | Groups, such as villages, schools, or hospitals, are randomized rather than individual participants. |
| Community trial / field trial | An intervention is tested in a community or healthy population, often for prevention. |
3. Secondary research designs
These evaluate already published evidence rather than recruit new participants.
- Narrative review: non-systematic summary of literature.
- Systematic review: structured, reproducible search, selection, appraisal, and synthesis of studies.
- Meta-analysis: statistical pooling of results from comparable studies, often performed within a systematic review.
- Clinical practice guideline: recommendations based on systematic evidence review plus assessment of benefits, harms, and feasibility.
Quick classification diagram
Research studies
│
├── Primary studies
│ ├── Observational
│ │ ├── Descriptive: case report, case series, cross-sectional, ecological
│ │ └── Analytical: case-control, cohort
│ │
│ └── Interventional
│ ├── Randomized controlled trial
│ └── Non-randomized / quasi-experimental trial
│
└── Secondary studies
├── Systematic review
└── Meta-analysis
High-yield comparison
- Cross-sectional: “What is present now?” Measures prevalence.
- Case-control: “What exposure may have led to this disease?” Best for rare outcomes.
- Cohort: “Does exposure lead to disease over time?” Best for measuring incidence and risk.
- RCT: “Does an intervention cause improvement?” Strongest primary design for treatment effects.
A concise overview of these categories and their uses is available in this study-design review.
This is a shared conversation. Sign in to Orris to start your own chat.