Skip to content
All articles

Statistics

Study Designs in Medical Research Explained: Cross-Sectional, Case–Control, Cohort and RCT

The study design you choose decides what questions your research can answer. Descriptive studies (case reports, case series, cross-sectional surveys) describe disease frequency and generate hypotheses.

Resora Research Team · 14 Nov 2026 · 4 min read

The study design you choose decides what questions your research can answer. Descriptive studies (case reports, case series, cross-sectional surveys) describe disease frequency and generate hypotheses. Analytical observational studies test them: case–control studies start from the disease and look back at exposures, and cohort studies start from the exposure and follow people forward to the disease. Randomised controlled trials (RCTs) assign the intervention by chance and give the strongest evidence of cause and effect. Cross-sectional studies give prevalence, case–control studies give odds ratios, and cohort studies and RCTs give incidence and relative risk. Pick the design that fits your objective, the rarity of the disease, time and budget, and ethics.

How study designs are classified

Start with one question: did the investigator assign the exposure?

  • No (observational): the investigator only observes. Descriptive and analytical studies belong here.
  • Yes (experimental): the investigator allocates an intervention, as in a randomised trial.

Among observational studies, ask a second question: is there a comparison group? No comparison group means a descriptive study; a comparison group means an analytical study.

Observational designs

Case report and case series

A detailed account of one patient or a group of patients with a similar condition. They have no comparison group, so they cannot prove cause, but they are often the first signal of a new disease or drug reaction.

Cross-sectional study (prevalence study)

Exposure and outcome are measured at the same point in time in a defined population.

  • Example: a survey of 800 adults in an urban ward measuring blood pressure and salt intake together.
  • Gives: prevalence, and an odds ratio or prevalence ratio.
  • Strengths: quick, cheap, good for planning health services.
  • Limits: cannot tell which came first, so it is poor for cause; unsuitable for rare diseases or short-lived conditions.

Case–control study

Start with people who have the disease (cases) and a comparable group without it (controls), then look back at past exposures.

  • Example: 150 women with breast cancer and 150 matched controls, compared for past use of a risk factor.
  • Gives: the odds ratio. It cannot give incidence.
  • Strengths: efficient for rare diseases and diseases with long latency; relatively cheap and fast; can study many exposures.
  • Limits: prone to recall bias and selection bias, especially in choosing controls; only one outcome per study.

Cohort study

Start with exposed and unexposed people, all free of disease, and follow them over time to see who develops the outcome.

  • Prospective cohort: the follow-up happens after the study begins. Retrospective (historical) cohort: records already exist, so exposure and outcome have both occurred, and the investigator reconstructs the follow-up.
  • Example: following 5,000 smokers and 5,000 non-smokers for 10 years for lung disease.
  • Gives: incidence, relative risk, and attributable risk.
  • Strengths: correct time sequence (exposure before outcome); can study many outcomes of one exposure; good for rare exposures.
  • Limits: expensive and slow; loss to follow-up; unsuitable for rare diseases.

Experimental designs

Randomised controlled trial (RCT)

Participants are randomly allocated to an intervention or a control group and followed for the outcome. Randomisation spreads known and unknown confounders evenly between the groups, which is why an RCT gives the most reliable evidence of cause and effect.

Key features:

  • Randomisation with concealed allocation, so nobody can predict the next assignment.
  • Blinding: single (participant), double (participant and assessor) or triple (plus the analyst).
  • Control: placebo or standard treatment.
  • Intention-to-treat analysis: analyse everyone in the group they were assigned to.

Limits: costly, slow, ethically impossible for harmful exposures (you cannot randomise people to smoke), and trial populations may not match real patients.

Other experimental designs

  • Quasi-experimental (non-randomised) trial: an intervention is given but allocation is not random, for example one village receives a health education programme and a similar village does not.
  • Crossover trial: each participant receives both treatments in sequence, with a washout period between.
  • Community trial / cluster RCT: whole communities or schools, not individuals, are randomised.

Comparison at a glance

DesignStarts fromTime directionMain measureBest forMain weakness
Cross-sectionalA populationOne point in timePrevalenceBurden of disease, planningNo time sequence
Case–controlDisease statusBackwardOdds ratioRare diseasesRecall and selection bias
CohortExposure statusForwardIncidence, relative riskRare exposures, many outcomesCost, loss to follow-up
RCTRandom allocationForwardRelative risk, risk differenceTesting an interventionCost, ethics, generalisability

How to choose a design

  1. Is the aim to describe or to test a hypothesis? Describe: cross-sectional or case series. Test: analytical or experimental.
  2. Can you ethically assign the exposure? If yes and resources allow, an RCT. If no, an observational design.
  3. Is the disease rare? Use case–control.
  4. Is the exposure rare? Use cohort.
  5. How much time and money do you have? Cross-sectional and case–control are fastest and cheapest.

The evidence pyramid

From weakest to strongest: expert opinion, case reports and series, cross-sectional studies, case–control studies, cohort studies, RCTs, and at the top systematic reviews and meta-analyses of RCTs. A well-done cohort study can still beat a poorly done RCT, so quality matters as much as design.

Keywords

study designs in research, types of study design, cross-sectional study, case control study, cohort study, randomised controlled trial, RCT, observational vs experimental study, prospective vs retrospective cohort, descriptive vs analytical study, evidence pyramid, how to choose study design, study design for thesis, epidemiological study designs, community medicine

References

  • Park K. Park's Textbook of Preventive and Social Medicine — chapters on epidemiology and epidemiological methods.
  • Hulley SB, Cummings SR, et al. Designing Clinical Research. 4th ed. Lippincott Williams & Wilkins; 2013.
  • Gordis L. Epidemiology. 5th ed. Elsevier; 2014.
  • Kirkwood BR, Sterne JAC. Essential Medical Statistics. 2nd ed. Blackwell Science; 2003.
  • Schulz KF, et al. CONSORT 2010 statement. BMJ. 2010;340:c332.

Related services