Assessing Attrition Bias -- Addendum

Attrition occurs when members of the initial research sample are not part of the final analysis sample, such as due to missing data or leaving the study. Both the overall sample attrition and the differences in attrition between the groups can affect the statistical equivalence of the sample and create potential for bias. The WWC has given careful thought to the ways in which attrition in randomized controlled trials can cause treatment and control groups to be dissimilar and possibly affect study findings. The WWC has determined how much attrition is "too much" by using a researcher-developed model that considers the effects of both overall and differential attrition. The attrition white paper (version 2.1) describes the model of attrition bias that informs the current WWC standard against which studies are assessed. It describes the working assumptions about the correlation between response and outcomes and examines whether the model parameters are generally consistent with data from three randomized trials. This attrition white paper addendum (version 3.0) provides additional empirical information and sensitivity analyses that support the use of the existing attrition bounds articulated in the white paper. [For "Assessing Attrition Bias" (version 2.1), see ED610408.]