1. In Example 14.1, adapt the procedure suggested by Hedeker and Gibbons (1997) to obtain populationwide estimates of the fixed effects (Intercept, Time, Drug, and Drug × Time), averaging over the dropout and completer groups. This involves weights based on the relative sizes of the totals completing (335) and dropping out (102). Note that MCMC avoids the need for delta methods to obtain standard errors on these pooled effects.