For statistical purposes, a "cohort" is just an identified group that we are going to investigate. The term comes from its use in the Roman military to designate a fighting unit composed of about 480 soldiers. We no longer assume that a cohort is of that size, but we do assume that we will treat the cohort as a group.

In general, we can indentify a cohort and

• get measurements on them now, follow them for some time getting more measurements in the future, and then do an analysis of the changes that have taken place: this is a propective study;
• get measurements on them now and do the study on those measurements: this is a cross-sectional study; or
• get measurements on them now, and look back in time to see if we can find differences in the past for them: this is a retrospective study.