Cohort and case-control studies are important types of study design in clinical and epidemiological researches. Although they are both methods to study associations between an exposure and a disease, their core logic is completely different.In order to evaluate the association between an exposure and an outcome, cohort studies group the populations by the research factor (exposure or not), and compare the event rate of the outcome in exposure and control groups through follow-up. Cohort studies are sequential from cause to effect. On the contrary, case-control studies derive cause from effect. Case-control studies group the populations by the outcome (occurrence or not), and compare the distributions of the exposure in case and control groups through retrospective investigation or collection of exposure information from historical records. This paper declares the logic that should be followed in statistical analysis of these two types of studies, as well as the specifications that should be followed in formulating statistical analysis plans and reporting results. The aim of this paper is to enhance the scientific nature of the research and improve the readability of the literature.

Cohort studies; Case-control studies; Specification; Statistical analysis; Logic