topic-distinct models??that enable a full specification with the distribution of the result by making use of random consequences to seize within just-topic correlations. The selection regarding which approach to choose partly depends on the intention on the analysis and the desired interpretation of your estimated consequences (population-average c