This is p. 168
organizations that congealed into a national organization (i.e., territorial penetration or diffusion); (2) whether the party was sponsored by an existing institution or arose on its own (i.e., externally or internally legitimated); (3) whether or not its founder stamped the organization with his charisma. Panebianco theorized that territorial penetration, internal legitimation, and the absence of charisma all predict to strong party institutionalization.
Institutionalization has a more interesting role in party theory as an independent variable. This is its prime function for Panebianco, who contended that "parties can be distinguished primarily according to the degree of institutionalization they attain" (1988, 55). For Panebianco, institutionalization not only predicts to organization of internal groups, but it also is a major inhibitor of party transformation or change (p.265). That is, highly institutionalized parties are more resistant to change. Moving from individual parties to the party system, Welfling theorized that the overall level of party institutionalization inhibited social conflict and promoted stability (1973, 54-58), and Dix (1992) viewed party system institutionalization as contributing to democracy in Latin America.
Comparative research on political parties pays great attention to parties' positions on issues with cross-national significance. Many such issues have been subsumed under the concept of ideology, but there concern has also been given to issues that do not fit common ideological concepts. As Duverger demonstrated forty years ago, ideology has a central role in party theory, primarily as an independent variable that affects other party characteristics. He spoke primarily about ideological types--communist, socialist, center, conservative, and fascist--but implicitly viewed them at positions along a left-right scale.
The left-right dimension: The theoretical centrality of ideology in party theory is demonstrated by the outpouring of empirical research on comparative party ideologies--most of which has been done by European scholars (Mavrogordatos 1987, 335). Since Duverger, scholars have advanced beyond the simple typologies of parties as communist, socialist, and so on, while accepting the principle of a single left-right dimension. Laver and Schofield (1990, 51-52) identified several methods used to order parties along a left-right continuum, including old-fashioned reading of primary and secondary sources (Taylor and Laver 1973, de Swaan 1973, and Dodd 1976), systematic surveys of experts (Castles and Mair 1984), content analysis of party platforms (Budge, Robertson, and Hearl 1987), and dimensional analysis of mass survey data (Huber 1989). Laver and Schofield conveniently summarize much of this research for 18 countries in an appendix (pp.245-266).
The survey by Castles and Mair (1984) deserves discussion. They asked over 100 western political scientists to classify parties in 17 countries on an eleven-point scale, ranging from Ultra-Left (0) to Ultra right (10). A total of 119 parties were rated, each by at least three country experts. Then Castles and Mair averaged their ratings to assign a left-right scale position to every party. They noted that "such a procedure necessarily does some damage to a multifaceted reality in which two or more political dimensions coexist and cross-cut each other" (p. 75). Nevertheless, they concluded that the parties' ideologies were judged by "general standards rather than purely national considerations" (p. 83), supporting the parties' placements on a single continuum.
Other cross-national research on party ideologies has uncovered at least two dimensions underlying parties' issue positions. One study used library research to code 150 parties in 53 countries on thirteen different issues commonly identified with the left-right continuum, and then analyzed their intercorrelations to disclose the latent factors. Eleven of the issues came out on two distinct factors (Janda 1980b, 147-149). Seven mostly economic variables correlated with what was called a "Marxist" scale, and four "civil rights and liberties" variables correlated with a "liberalism" scale.
A "two-factor" interpretation of parties' issue positions was also supported in the most ambitious effort at analyzing party ideology. Budge, Robertson, and Hearl (1987) reported the results of a project sponsored by the European Consortium for Political Research to perform a content analysis of the election manifestos of all significant parties in 19 countries from 1945 to 1983. Rather than judging the political substance of the parties' statements, the group of country experts engaged in the more objective but controversial procedure of counting the number of sentences in the platform according to 54 categories in seven broad policy domains. "For this implies that the most important aspect of the documents is the degree of emphasis placed on certain broad policy areas, rather than each party's support for, or opposition to, a specific policy within these areas" (p. 24). After factor analysis, this method of content analysis produced two major factors: "a central and clearly Left-Right cleavage in most of the countries under consideration--15 out of 20" (p. 392) and a second dimension that "often seems to reflect Left-Right contrasts in a modified form" (p. 395). So despite the possibility of great complexity in parties' issue positions across nations, existing research suggests that the variation can be accounted for with only