Reanalysing data from a former empirical study the GOODMAN approach for analysing nonmetric data is compared with a method of weighted regression analysis by GRIZZLE, STARMER, and KOCH, which recently has been transformed into a neat computer program greatly facilitating its application. By describung both approaches as modfied regression analyses common traits and differences can be pointed out very clearly. It turns out that in the majority of research problems the GSK approach is superior due to the fact that substantive results will both be more detailed and entail a higher degree of intuitive appeal.