Goodman (1979) and Andersen (1980) recently presented a log-multiplicative model for the analysis of association between two discrete variables. This model can be viewed as a log-linear model with certain restrictions on the interaction parameters. The usual log-linear interaction parameters are factorized in two sets of parameters, one set pertaining to the row variable and one set pertaining to the column variable. These two sets of parameters can be viewed as row and column scores. In this article the model formulated by Goodman and Andersen is reviewed and some generalizations for multiway tables are discussed. A three-way table is analyzed as an example.