This is p. 174
congruence among party members or activists. It is much easier to study coherence in party behavior, which is more open to inspection.
Cohesion: Party cohesion, defined as the extent to which parties vote together in legislative bodies, is readily amenable to study and legislative voting has generated a huge literature in the United States. Despite the publication of such voting data in foreign countries, research on party cohesion is less common in cross-national analysis, although there are many studies in individual countries (Collie 1985). There has also been surprisingly little research on the effect of organizational attributes on party cohesion--despite Duverger's early hypotheses. He said that "domination over the parliamentary representatives by the party" was due to the "general structure of the party and its general orientation" (1963, 202). His argument can be reformulated into three hypotheses: the more centralized the party, the higher the cohesion; the greater the leftism, the higher the cohesion; and the more ideologically extreme, the higher the cohesion.
Ozbudun (1970) conducted the most concerted effort at the comparative study of party cohesion. He and other scholars (Turner and Schneier 1970 and Loewenberg and Patterson 1979) concentrated on six environmental factors in explaining party cohesion: presidential government, federalism, multiple parties, ideological polarization, single-member districts, and legislative effectiveness. Harmel and Janda (1982) found that 32 percent of the variance in legislative cohesion could be explained with only two variables (legislative-executive structure and legislative effectiveness). Their analysis supported Epstein's contention that separation of powers was the key variable in explaining the low cohesion of American parties compared with those elsewhere (1980, 315-350).
Factionalism: There is another aspect to coherence in party behavior, factionalism. In a pioneering article, Zariski (1960) defined a faction as "any intra-party combination, clique, or grouping whose members share a sense of the common identity and common purpose and are organized to act collectively--as a distinct bloc within the party--to achieve their goals" (p.33). Sartori (1976,76-79) proposed a typology of factions based on their interest (power or spoils) and their principles (ideologies or ideas). Another approach distinguished between factionalism based on ideology, issues, leadership, and strategy (Janda 1980b). Over 100 parties were scored on each type of factionalism on a seven-point scale. The ideological basis was somewhat more common than the others, but all types of factionalism tended to be interrelated; e.g., if a party had ideological factions, it also tended to have leadership factions.
Most research on factionalism has regarded it as a dependent variable in party theory. Belloni and Beller (1978) edited a collection of studies that sought to describe and explain factionalism. It included an article by Zariski (1978), who listed a "dozen most tenable and agreed-upon generalizations on factionalism" (p.32). Most of his propositions linked factionalism to environmental factors, such as the nature of the electoral system. Nevertheless, Belloni and Beller concluded that factionalism was also due to the "sociological complexity of the party; ideological looseness of the party; the origin of the party in a merger of predecessor parties [i.e., its conditions of origin]; [and] the party's internal looseness or decentralized structure" (p.435).
The Belloni and Beller book largely avoided regarding factions as independent variables. However, party factionalism is an important cause of low voting cohesion and needs to be included in any theory of party government. Factionalism's effect on party cohesion is apparent in American politics. Southern and Northern Democrats often opposed each other in Congressional voting until an ideological realignment began to occur in the South around 1970. Since then, party cohesion (as measured by Congressional Quarterly party unity scores), has almost steadily increased. Sinclair attributes this rise in voting cohesion to decreased ideological heterogeniety of the Democrats (1990, 241-242). Finally, factionalism and cohesion together (the general concept of coherence) are critically important in coalition theory, which--with few exceptions (Luebbert 1986 and Laver and Shepsle 1990)--assumes that parties act as units (Laver and Schofield 1990, 17-22). Brady and Bullock (1985) provided a general review of literature on parties and factions in legislatures from the standpoint of legislative behavior. They saw "much to be done" in linking the legislative party to external components and in investigating the relationship "between the degree of fractionalization and/or factionalization of legislative parties and the way in which parties distribute task and power" (pp. 175-176).
In a major section of his book, Duverger (1963) discussed at length the concepts of party membership, degree of participation, and nature of participation. Individually, these topics have drawn a great deal of attention in the parties literature, but there has been only mixed success in integrating the discussion. I prefer to subsume these topics under the concept of involvement,