This is p. 173
meaningful distinction, for parties (like British Labour) with indirect membership are less autonomous than those with only individual members. Nevertheless, fewer than 10 percent of the world's parties in 1960 had any indirect members, while about 20 percent had no formal membership requirements (Janda 1980b, 93). Because party membership data are unreliable and often not applicable, Lawson advised paying more attention to the number of votes actually received by a party than the number of members it claimed (1976, 93). Alternatively, one can study party identification or "attachment" as a form of party affiliation (Richardson 1991). In the case of Europe, this research is facilitated by two decades of Eurobarometer surveys, which permit longitudinal, comparative research, albeit with some problems (Katz 1985). In several countries from 1974 to 1988, Schmitt (1989) found a decline in people who felt "close to" a party, but the pattern tended to be mixed elsewhere, and Germany even showed a slight increase of attachment. Formal party membership is discussed below under the concept of involvement.
Source of leaders: Panebianco uses the term, "dominant coalition," for the group of leaders who control the principal power resources of a party organization (1988, 38). In slightly more than 25 percent of the parties in 1960, most leaders came from one sector of society, indicating a lack of autonomy (Janda 1980b, 94). Lawson (1976) analyzed diversity in party leadership in three countries, but this topic has not been studied heavily through comparative research. Everyone seems to agree on the importance of leadership, particularly in party change (Wilson 1980, 542-544), but it is notoriously difficult to study rigorously.
Relations with Domestic Parties: Duverger identified three types of cooperative interparty "alliances": electoral, parliamentary, and governmental (1963, 330-351). Electoral alliances infringe the least on autonomy, and governmental alliances the most. "Satellite" parties of ruling parties in Eastern Europe demonstrated the most severe loss of autonomy (Fischer-Galati 1979), but this subject drew little attention from party scholars. Although there are other dimensions to research on party alliances (Gilberg 1989, Kolinsky 1987), the emphasis has been on governing coalitions in parliamentary democracies (Pridham 1986, Laver and Schofield 1990). The coalition literature, which relies heavily on rational choice theory, is reviewed below under party strategy. The concern here is whether coalition possibilities limited party autonomy, a concern that surfaced in the concept of a "relevant" party. Sartori said that a party is relevant if it "affects the tactics or party competition" by having "coalition potential" or having "blackmail potential"--possessed by an antisystem party or one that can veto political arrangements (1976, p.123; see also Herzog 1987). Budge and Keman theorized about the effects of antisystem parties on party behavior (1990, 44), while Hofferbert and Klingeman (1990) contended that the FDP in Germany acts as a swing party due to its blackmail potential.
Relations with Foreign Organizations: Parties can have relationships (mostly but not exclusively) with international party organizations or foreign governments that can impinge on their autonomy. This was clear in the case of communist parties but it also applied to Socialists (Pelinka 1983, 108-124), Christian Democrats (Irving 1979), and Liberals (Hrbek 1988)--all of whom maintained international organizations and demonstrated varying degrees of solidarity in international bodies (Pridham and Pridham 1981). Goldman (1983) has been the foremost scholar of such "transnational" parties, and the topic is certain to grow in importance with the growth of the European community. For example, Eijk and Oppenhuis (1991) analyzed the competition for votes among parties competing in direct elections to the European parliament in 1989.
The concept of coherence was introduced into comparative politics by Huntington (1965, 403-405) and developed by Anderson (1968, 396-397) for the study of state and local parties. In comparative analysis, coherence has been defined as "the degree of congruence in the attitudes and behavior of party members" (Janda 1980b, 118). "Behavior" is emphasized because few cross-national studies have surveyed the attitudes of party activists or members--as opposed to voters. One exception was the European Political Parties Middle-Level Elites Project, which collected questionnaires from activists attending more than 65 national party conferences in nearly all the European community countries (Niedermayer 1986). Most research from this project was published in German, with a typical study focusing on the representation of social groups within the party elites (Niedermayer and Schmitt 1983). This mammoth survey of middle-level party elites was criticized by Pierre (1986) for being done in isolation from other aspects of party organization. In any event, there are few such studies (but see Dalton (1985) below) because massive research is required to compare attitudinal