identify parties for study. The first restriction recognizes that we are interested only in parties that operate in national politics, which excludes some local parties. The second requires that the parties achieve a given level of importance in national politics during our time period (1950-1962), defining importance in terms of strength among the population and stability of existence. These criteria are embodied in the specific operational definitions used to identify legal and illegal parties for inclusion in our study.

For legal parties, our operationalization requires that the party win at least 5 percent of the seats in the lower house of the national legislature in two or more successive elections. The seat requirement and the successive elections requirement are intended to certify the party's strength and stability within the political system.

For illegal parties, our definition is not so easily specified. Conceivably, "size of membership" could be incorporated into an operational definition, but membership data on illegal parties (or on legal parties, for that matter) are difficult to obtain and virtually impossible to verify. The concepts of strength and stability of an illegal party are specified instead in terms of support of a sizable proportion of the population, at least 10 percent, which is sustained over a certain period of time, at least five years. The 10 percent and five-year criteria are designed to reflect in a roughly comparable way the 5 percent and two successive elections criteria for legal parties. In applying the operational criteria for illegal parties, country analysts had to draw inferences from the literature and exercise considerable judgment in identifying parties for inclusion in the project.

By limiting our attention to political parties that demonstrate a certain level of strength and stability during our period of interest, we (1) reduce the number of parties to a more manageable level of hundreds, rather than thousands, of units to study, (2) simultaneously focus our research on parties for which information is more likely to be available, and--most important--(3) include those parties, both legal and illegal, that are most likely to have measurable consequences upon national governments. The complete list of parties in the ICPP project is given in Table 1.3. The first number in the three-digit party identification code refers to the cultural-geographical region of the world in which the party operates. The first and second numbers together identify the country itself and correspond to the identification numbers in Table 1.1. The third number has been arbitrarily assigned to specific parties within a country.

As we dealt with a great variety of literature on party politics across the world during the bibliographic and indexing phase of the project, I began to detect a relatively small number of basic concerns that appeared to underlie most of the specific observations that scholars were making about parties. By 1969, I came to identify eleven major concepts that seemed to encompass most writers' basic concerns. These eleven concepts became the foundation of the conceptual framework of the ICPP project. A twelfth concept was added later to provide a reference for two of the original eleven.¹ I contend that these concepts capture most of the major dimensions of variation of political parties across nations.

The major concepts in the ICPP conceptual framework can be separated into those that pertain to a party's external relations and those that relate to its internal organization. The analysis of a party's external relations is judged according to eight concepts: institutionalization, governmental status, social attraction, social concentration, social reflection, issue orientation, goal orientation, and autonomy. Its internal organization can be analyzed according to the remaining four concepts: degree of organization, centralization of power, coherence of behavior, and involvement of its members.

The conceptual framework of the ICPP project consists of these major concepts plus approximately 100 "basic variables" grouped into twelve corresponding "clusters" of variables. Each basic variable within a cluster constitutes a specific indicator of the more abstract concept which subsumes it. In most instances, the basic variables in these variable clusters are proposed as alternative indicators of the concept represented by the cluster. Under the assumption that these indicators are all adequately intercorrelated, they can be combined in an "additive" approach to concept measurement.

Of course, the extent of intercorrelation among indicators within a cluster is a matter for empirical study. Chapter 14 reports the results of such an analysis. Although the assumption of high intercorrelations among the proposed indicators of a given concept was not completely substantiated in all instances, the empirical results gave strong support to the initial conceptualization. Whereas parts of the original conceptual framework were modified (see Chapter 14), most of it was validated by the strong correlations observed among the indicators. Notwithstanding the several conceptual modifications dictated by the data analysis, the basic variables are discussed in Chapters 1 through 13 according to their placement in the original conceptual framework. This decision enables the reader to ascertain the intellec-

---

1. As explained in Chapter 14, the conceptual framework underwent a slight reorganization after the twelfth concept was added. That last concept, social reflection, plus the concepts of social attraction and social concentration were combined under the more general concept, social support. The result was a framework of ten rather than twelve overarching concepts. This reorganization was mainly a matter of tidying the conceptual structure and had no major substantive consequences. Therefore, no further note will be taken of this modification until Chapter 14, when the original conceptual framework is tested with data collected from the study.

go to page 8