A Weight of Evidence Account

Abstract

In this chapter, I explicate the weight of evidence account. The weight of evidence accorded a hypothesis is highly dependent on the accuracy of the evidence, which is understood to mean a combination of validity and precision. Validity is concerned with biases and confounding, which can distort the measure of effect, whereas precision is often associated with the width of confidence intervals around the effect measure. Evidence considered to be of greater accuracy is accorded greater weight. The weight of evidence notion can be applied to single observations or studies, or groups of these. For therapeutic hypotheses, the randomized clinical trial affords the greatest degree of accuracy. Threats to validity in etiological studies are greater, sometimes resulting in lower weights being accorded. Maximum weight of evidence can be achieved for diagnostic hypotheses by meeting a “gold standard.” The processes by which evidence is determined and quantified are considered, and the ways in which evidence is aggregated, evaluated, and applied to problems including determining best therapy or improving public health policymaking are discussed.