V.O. Key's works Iiave piayed a centrai rde in the deveiopment of comparative state pditicai anaiysis. in Soutiiem Politics (1949) he posed a question of enduring interest and importance: how does ciass conflict and its reflection in party divisions affect public pdicy? He argued that the organization of the electorate into differing, enduring, and cohesive factions or parties creates issue oriented pditicai dialogues, whk:h affect the interests favored in the pditicai process. His argument remains one of the most intriguing in the discipline.When state analyses in the 1960s moved to utiiize regression methods, his hypotheses dominated inquiry. This transition in methods generaiiy was viewed as essentiaiiy a shift in anaiytic techniques, and not a movement away from Key's concerns or approach. The argument of this essay is that these eariy anaiyses (Ranney and Kendaii, 1954;Schiesinger, 1955;Dawson and Robinson, 1963;Dye, 1966) and subsequent ones (Lewis-Beck, 1977; Tucker, 1982; King, 1986) do not capture Key's concems about parties. We stiii do not have valid assessments of the impact of pditicai parties on pubiic pdicy.The difficuity stems from the deveiopment of inter-party competition indexes tfiat attempted to capture Key's kJeas of electoral organization. Key focused on the effect of eiectorai aiignments within states on state pdicy. He carefuiiy expiored t}ehavior in states before characterizing them. Subsequent studies discarded a concern for these within-state relationships and focused oniy on state-ievei aggregates and their "effects." Key's notions of the minimai requirements of comparison were not utiiized by subsequent researchers. Anaiysts ignored Key's repeated and explicit warnings that such characterizations were uniikeiy to be appropriate to capture the matters he was examining. The change, in retrospect, has been a profound one. Anaiyses are conducted as if they assess Key's questions, but the crucial index reiied on is invaikJ. The comjjarative approach that Key deveioped wiil first be reviewed. The aitemative approach and its deveiopment wiil then be 246 Policy Studies Journai modei for comparative anaiyses (Burnham, 1986:32-33). Despite Key's anaiysis of multiple states, in the 1950s the predominant approach remained singie state case studies.T he effort to move to "comparison" began in 1954 with the characterizatkin of state inter-party competition conditions by Ranney and Kendaii (1954). Various efforts fdiowed which suggested improvements in the index (Sciiiesinger, 1955; Gdembiewski, 1958; Hofferijert, 1964; Ranney, 1965; Pfeifter, 1967; DavkJ, 1972; Broh and Levine, 1978; Wright, Erickson, and Mciver, 1985). These indexes mari