Ask Mallard

Cohort or case-control?

Choose how to sample people and observations for an etiologic question.
NHMRC EVIDENCE LEVEL · 2009Not yet classifiedProvisional design classification. Study quality and certainty of findings require separate appraisal.
Why this level?Specify the research question: intervention, diagnostic accuracy, prognosis, aetiology, or screening intervention. Other question types are outside this hierarchy. NHMRC 2009 source

This page compares sampling designs, not a single study. The research question, sampling scheme and actual study conduct determine the applicable evidence classification.

The decision

Start with the population, exposure contrast and outcome horizon. A cohort identifies an eligible population and follows or reconstructs its outcomes. A case-control study samples cases and controls from a defined source population, then measures or retrieves exposure histories. Either can use existing records. Retrospective does not mean case-control.

When a cohort is useful

A cohort is attractive when the eligible population and follow-up can be enumerated, exposure data are available across it, and several outcomes or absolute risks matter. It can establish a denominator for risk or person-time rates if ascertainment is adequate. A rare exposure can be an efficient starting point. Long follow-up and expensive exposure measurement may make a full cohort difficult.

When case-control sampling is useful

Sampling cases and suitable controls can reduce measurement costs when the outcome is rare or exposure assessment is expensive. A nested case-control design can retain a clear cohort source and time origin while measuring stored specimens in a subset. Specify the sampling scheme before analyzing the selected records.

What changes in the analysis

  • Controls should represent the exposure distribution in the source population or risk sets specified by the sampling design; they are not simply people with no diagnosis in an unrelated clinic.
  • With incidence-density sampling, the odds ratio estimates an incidence rate ratio without requiring a rare outcome. With cumulative sampling, a rare-outcome approximation has a different role. State which design you used.
  • A sampled case-control dataset alone does not usually supply absolute risk. Recovering it requires the relevant sampling fractions or additional population information.
  • Matching changes both sampling and analysis obligations. Preserve the matching structure and justify the matching variables; do not size matched work with an unmatched formula.

One population, sampled three ways

The same six people shown three ways. As a cohort, everyone is followed and every outcome counted. With cumulative case-control sampling, controls are chosen at the end from people who never had the outcome. With incidence-density sampling, controls are chosen from people still at risk when each case occurs, so a control can later become a case.Cohort: everyone followed, every outcome countedPerson APerson BPerson CPerson DPerson ETimeCase-control, cumulative sampling: controls chosen at the end, from people who never had theoutcomePerson APerson BPerson CPerson DPerson ETimeCase-control, incidence-density sampling: controls chosen from people still at risk wheneach case occursPerson APerson BPerson CPerson DPerson ETimePerson D is a control for the first case and becomes a case later. Which scheme was used decideswhat the odds ratio estimates.Outcome occurs (a case)Sampled as a control
Conceptual diagram. It shows the structure of the idea, not data from any study.

Example

A biobank has follow-up on all eligible participants but an expensive assay is needed. A cohort analysis using an assay already measured for everyone and a nested case-control analysis using assays from sampled risk sets are different feasible designs. Choosing between them requires assay cost, case count, sampling and the desired effect measure.

Record in the workspace

For this aim, record the source population, eligibility date, exposure ascertainment, case definition, control-sampling rule, matching and planned effect measure. Keep the choice open only while both sampling strategies remain feasible.

What to watch out for

Control sampling determines what a case-control odds ratio estimates. A convenient control group cannot be repaired by a more complicated regression.

How to report it

STROBE is the guideline for both designs this page compares, with a separate checklist for cohort and for case-control studies. Report which design was chosen and why, the source population either way, and use the checklist for the design actually run rather than a combined one. A reporting guideline is not a design manual: use it to make the study transparent, not as a substitute for careful design. Check the target journal's own requirements as well, which this page cannot know.

Recent studies that used this
Searching PubMed…
Related Learn pages
Sources & further reading
  1. Case-control studies: Increasing scientific rigor in control selection (opens in a new tab) · Res Pract Thromb Haemost · doi 10.1016/j.rpth.2023.100090
  2. Observational studies: cohort and case-control studies (opens in a new tab) · Plast Reconstr Surg · doi 10.1097/PRS.0b013e3181f44abc
  3. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies (opens in a new tab) · BMJ · doi 10.1136/bmj.39335.541782.AD
Prepared with AI assistance. Methodological and source checks are documented in this release; this is educational material, not independent professional approval. Reporting guidance supports transparent reporting and does not establish validity.
Last source review: 2026-09-09
AI-generated educational content
Mallard uses reputable methodological and reporting sources, but this material may contain errors or omit important context. Verify important methodological decisions with the cited sources and, when appropriate, a qualified research methodologist or statistician.