The Importance of Accounting for Correlated Observations

Correlated data arise when pairs or clusters of observations are related and thus are more similar to each other than to other observations in the dataset. Observations may be related because they come from the same subject—for example, when subjects are measured at multiple time points (repeated measures) or when subjects contribute data on multiple body parts, such as both eyes, hands, arms, legs, or sides of the face. Observations from different subjects also may be related—for example, if the dataset contains siblings, twin pairs, husband-wife pairs, control subjects who have been matched to individual cases, or patients from the same physician practice, clinic, or hospital. Cluster randomized trials, which are performed to assign interventions to groups of people rather than to individual subjects (for example, schools, classrooms, cities, clinics, or communities), also are a source of correlated data because subjects within a cluster will likely have more similar outcomes than subjects in other clusters.