A case-control study’s group sizes are chosen for comparison, not a population risk estimate

Ask how people entered the two groups

A table with an exposure and outcome can come from different designs. Locate whether investigators began with exposure groups and observed outcomes, or selected outcome-defined cases and controls for comparison. The timeline of the published paper is not enough. Records can be analyzed later while the underlying selection logic remains the important distinction.

CDC’s archived analytic-epidemiology lesson describes case-control selection by outcome status and comparison of earlier exposures. We use that design concept without importing its disease examples or findings. This article does not evaluate any specific vaping exposure, disease association, quitting intervention or person’s health. It identifies the sampling question behind a reported number.

US Centers for Disease Control and Prevention: Principles of Epidemiology, Lesson 1 Section 7 (archived)

Use an invented recruitment ledger

Suppose an imaginary project recruits twenty case records and forty control records according to its plan. Twenty of sixty selected records are cases. That fraction describes the assembled study sample, whose case-control balance was shaped by recruitment. It does not establish that one third of the source population experienced the outcome.

If the same imaginary project had selected eighty controls instead, the sample’s case fraction would change without changing what happened in the source population. These invented values illustrate selection, not a real study or an effect calculation. The helpful output is a short ledger showing why each group was eligible, where it came from and how its size was determined.

Keep the exposure comparison on its proper scale

A case-control report may give an odds ratio and uncertainty, potentially from an adjusted model. Keep that measure’s actual name and direction. Do not use the selected sample’s case fraction as an unreported baseline risk for converting the ratio into a population probability. A design-specific interpretation can require methodological expertise beyond this introductory guide.

Controls should represent a relevant comparison for the source of cases, but a reader should inspect the authors’ actual selection explanation rather than inventing one. Also ask how earlier exposure was measured and whether its timing fits the question. Selection, recall and other design issues remain relevant even if the final table looks complete and its numerical estimate is precise.

Write what the study can and cannot count

A useful note says how cases and controls were defined and selected, what exposure contrast was estimated, and which population-risk information was not supplied. If recruitment details are unavailable, retain that gap. Do not label every case-control result invalid, or turn a design limitation into proof that an association is absent.

For any medical decision, bring the original paper to a qualified professional. A personal vaping log cannot fill the source-population denominator missing from a research summary. Ordinary reading ends when the sample’s composition no longer becomes a false population forecast and the reported comparison remains attached to its genuine design, outcome definition and uncertainty.