A model was designed to specify the components of blood pressure (BP) in a reactivity study. The model considered six components for BP in response to tasks: the average resting BP across all subjects, a given subject's deviation from that average, the average task effect for a specific task, a given subject's deviation from the average task effect, the average effect of repeated assessments of BP within a given task, and an error term. The model and data from 71 adult men were used to estimate the components represented by averages. The variances and covariances of the components were represented by deviations from this average. Utilizing the likelihood ratio statistic with the bootstrap null distribution, the model gives a reasonable representation of the data. Alternative models were also tested; however, they fell short of representing the data well. The fitting and testing of components in models like ours may offer some guidance in the design of future reactivity studies.