Cohort studies are one of two kinds: (a) either they start off with a single group of people from whom standard data are collected and the outcomes years later are related to people with different characteristics found in this data collection; or (b) they start off with two or more groups of people (i.e., coal miners who work underground and miners who work on the surface) and follow up to see if they end up with the same or different outcomes. Cohort studies can be two types: (a) they either start in the present time and await outcomes in the future (prospective studies); or (b) the outcomes are measured in the present time on groups of people who can be categorized as having had certain characteristics in the past (i.e., a follow-up study of people born 20 years ago about whom a lot of data was collected at the time of birth to see if maternal complications have had a lasting effect on their offspring (retrospective studies)).

Cohort studies are not analyzed by odds ratio, but by RR (Relative Risk or Risk Ratio). The rate that an outcome occurs in the exposed group is divided by the rate in the unexposed (i.e., if smokers develop 80 lung cancers per 1000, and nonsmokers only 5 per 1000, then the RR = 80 divided by 5 = 16; smokers were 16 times more likely to develop lung cancer). Other analyses are also possible.