How are power calculations used in evaluating study feasibility?
For a study of environmental exposures that occurred in a specific community, we do not have the option of simply increasing the study’s size to gain statistical power of 80% or more. This is because the size of the population that is exposed to an environmental contaminant is already fixed. In our evaluation of the feasibility of conducting follow-up studies for various health outcomes for the Endicott area, we used the power calculations to make judgments about whether in-depth studies in the Endicott area alone would have sufficient statistical power. Because power estimates differ depending on study design decisions and assumptions, particularly about how successful the study will be at locating and recruiting eligible study subjects, we drew conclusions using the assumptions we thought were most appropriate and realistic. We attempted to be moderate in our estimates and assumptions, neither overly optimistic nor pessimistic.
