Cohort

Classic Research

A cohort, derived from the Roman word for a division within a legion, can be any kind of group or grouping of interest to the researcher. Classic cohort research such as the Framingham Heart Study involved a large number of participants given several measures and followed over time to determine changes in their health and how it related to other variables. Normally, prospective cohort studies rely on a large sample size that allows for loss of follow-up (dropouts). Not every member of the Roman legion returned from battle and not all of the participants involved in a study continue to participate over time. In addition, larger cohorts allow researchers to have more confidence in their findings. The Framingham study began with more than 5,000 individuals and sought to discover relationships between certain variables (e.g., smoking, body weight) and the outcome of a myocardial infarction (i.e., heart attack). Researchers were interested in determining the risk factors associated with a particular health outcome.

Another classic example of a cohort research design, the Nurses’ Health Study, also used a large sample. Here, the initial outcome of interest was cancer and researchers were interested in how it correlated with the use of oral contraceptives. The loss of follow-up was complicated by needing an adequate number of participants who took oral contraceptives and a sizable population of those who did not. One of the disadvantages of typical cohort studies is a need for large samples in order to achieve statistical significance and to adjust for loss of follow-up. However, this disadvantage is counterbalanced by the ability of cohort studies to observe phenomena over time. In fact, the Framingham Heart Study and the Nurses’ Health Study are ongoing longitudinal designs that continue to collect data from new generations of cohorts.

...