This is p. 171
well-developed organizational theory literature, the conceptualization of party structure has an ad hoc quality, with little attention paid to creating reliable measures of party organization (Janda 1983). In this case, the American literature on party organization has an earlier history (see Crotty 1968) and is better developed. Creative measurement of party organization is evidenced in the collaborative work by Gibson et al. (1983) and Cotter et al. (1984) on comparative analyses of state and county party organizations, and in later work by Gibson et al. (1989). This research found a surprising amount of organizational strength among state and local parties in 1980 and, if anything, increased levels of strength when the parties were studied again in 1984. Unfortunately, these studies were rooted in the American context and do not help much in comparative cross-national analyses.
Wellhofer (1979) is one of the few scholars who devised measures of a broad organizational concept for application to parties across nations. Wellhofer proposed the concept of "organizational encapsulation," defined as "the elaboration of party sub-units to envelop as many of the day-to-day life activities of the membership as possible" (p.206). Encapsulation was measured with three objective indicators of party membership and local organizations for socialist labor parties in Argentina, Britain, Norway, and Sweden. He used time series analysis in a longitudinal study to demonstrate how these organizational inputs were linked to vote outputs (see also Wellhofer 1981). Unfortunately, Wellhofer dealt primarily with that one organizational concept. On the other hand, Panebianco (1988), conceptualized party organization more richly--discussing its dominant coalition, institutionalization, and organizational complexity--but was less clear on how to measure these concepts objectively.
Most of the empirical comparative research on party organization again draws inspiration from Duverger. He advanced propositions using such concepts as direct/indirect structure, basic elements, organizational articulation, and centralization of power. Lane and Ersson thoughtfully modified and applied these concepts in their study of European parties (1991, 123-128). Janda (1980b) interpreted Duverger through the organizational theory literature to differentiate two major dimensions of organizational structure: degree of organization and centralization of power. Degree of organization, which is similar to Duverger's articulation, referred to the complexity of structural differentiation. Centralization of power, following Duverger, referred to the location and distribution of authority. When measured by two highly reliable scales over more than 100 parties, these two dimensions were empirically unrelated. Despite these findings, the distinction between organizational complexity and centralization of power have not been not widely recognized, which is why they are treated together in this section.
With few exceptions, the comparative parties literature has paid relatively little attention to conceptualizing party organization and even less to measurement issues. This is particularly true for the concepts of complexity and power. For example, Sartori (1976) simply talked of an undifferentiated "organization" dimension on which parties could be classified as "organized," "organizationless," or "half and half" (pp. 76 and 81). Lawson (1976) created a set of six types of organization based on the locus of power and involvement of active members (p.78). Indeed, von Beyme despaired over the complicated network of organizational influences in European parties and said, "A typology of similar cases is the utmost that seems possible" (1983, 367).
Concepts of organizational structure are also not prominent in empirical party theory, with Duverger again an exception. He used organizational variables as both independent and dependent variables. For Duverger, centralization of power was a cause of party cohesion (discussed below) and was caused by leftist ideology. He also explained party organization by the conditions of origin (e.g., whether it began inside or outside the parliament; early or late in the nation's political development). Several scholars (Hodgkin 1961 and LaPalombara and Weiner 1966) have used conditions of origin to explain party structure in the third world and European parties, while Koelble (1989) cited both ideology and conditions of origin in explaining the West German Green Party's decentralized structure. Few scholars have used centralization as an independent variable in comparative research. An exception is Dalton (1985), who combined a European voter survey with a simple measure of party centralization to show that the more centralized the party, the more accurately voters perceived its cues.
Complexity of party organization and centralization of power are also caused by environmental factors. In their study of 95 parties in 28 democratic and quasi-democratic nations, Harmel and Janda (1982) found that parties in the same country tended to be similarly organized. For complexity of organization, 44 percent of the variance was explained with six environmental variables (modernity, population size, electoral system, restrictions on suffrage, recency of democratic experience, and lack of party competition). For centralization of power, 35 percent of the variance was explained with six variables (country size, federalism, and aspects of legislative-executive structure).
As noted earlier, Eldersveld, Epstein, Keefe, and others have characterized both major U.S. parties as extremely decentralized. A quantitative cross-national analysis of more than 100 parties documented their claim: