sights and vigor into comparative politics generally. We need to adopt a holonational approach to the study of political parties to stimulate our thinking about the important dimensions of variation among political parties. We need to assemble comparable data on parties in a holonational framework to give empirical import to our conceptualizations and to test the theories that embody our concepts.
Not all scholars agree on the desirability of, much less the need for, a holonational approach to the comparative study of political parties. Indeed, one of the most consistent criticisms of my project that I encountered was that I ought not and even could not conduct a study of political parties that involved such different organizations as found during our time period in Nicaragua, Burma, Guinea, East Germany, Portugal, and France. My position on whether such parties ought to be compared should now be clear; I think it is essential for the development of party theory to study different as well as similar parties. On the other point, whether or not such parties can be compared, the issue can be settled only by empirical investigation.
This work is aimed at advancing that investigation. It focuses on 158 political parties operating in 53 countries during the time period 1950 to 1962, but it also traces the fate of those parties through 1978 and embraces 50 new parties that qualified for study between 1963 and 1978.
A word is in order about what this book is and what it is not. It was not designed as an encyclopedic compendium of facts about all the world's political parties written by experts in the politics of each country. Although the information it contains has value along that line and will undoubtedly be used for reference, the book was designed with a different purpose in mind. This work was intended to impose intellectual order upon the mass of facts about political parties--to create concepts for interpreting such facts and to harness them in the service of comparative analysis. Thus, the conceptual framework offered herein is as important as the facts amassed under it. Indeed, the data that issue from this study are the joint products of the conceptual framework and the factual information that it interprets. For example, the specific procedure used by a party for selecting its national leader is simply a fact until it is interpreted in the context of a continuum of variation in party practice that suggests the centralization of power within the party. Thus, we assign quantitative scores to each party for many "basic" variables (e.g., "selecting the national leader") that serve as indicators of a smaller number of broader concepts (e.g., "centralization of power") that are crucial to evaluating the character of political parties.
The statistical tables in Part One summarize the results of scoring a sizable sample of the world's parties on the basic variables in our conceptual framework. By consulting these tables, the reader can learn how the parties of the world tend to be distributed on important variables that pertain to key concepts in the comparative analysis of political parties. One can learn which practices are common, which are rare, and what each practice signifies in the analysis of party politics. For students of comparative politics, the main value of this work probably lies in these survey results. Serious researchers who wish to push the analysis beyond the simple distributions reported in the tables will want to work directly with the quantitative scores listed in Part Two. Their research can be greatly aided through the facilities of the Inter-University Consortium for Political and Social Research at the University of Michigan, which archives and distributes the quantitative data from the study. Part One, in effect, constitutes the codebook for the completely documented machine-readable data set available on computer tape from the consortium.4
The listing of party data in Part Two probably serves another set of readers--those who merely seek basic information about particular parties. They will be more interested in the encyclopedic function of the volume. It is hoped that the restricted coverage of the study will still be adequate to serve their interests as well.
basic concepts that form the core of the study's conceptual
framework appear broad enough to encompass the major
dimensions of variation of political parties as they
operated within ten cultural-geographical areas of the
world. The information available for scoring parties on the
numerous indicators of each of the twelve concepts varied
considerably in quantity and quality across countries.
Despite the constraints of the available information, we
succeeded in scoring most of our parties on most of our
approximately 100 indicator variables. We relied heavily on
modern microfilm and computer techniques of information
retrieval in our research, and we developed and employed
several new methods of data quality control to increase the
reliability of our data. We state explicitly whether our
judgment about the score for a party on a given indicator is
an inference from sketchy information or whether our
decision is based on consistent information from multiple
sources. Each coding judgment, moreover, is accompanied by a
note or longer comment elaborating the basis of the scoring
for that party. Because the information available for many
countries rarely anticipated the coding categories for our
variables, we tried to be as open and explicit as possible
in showing how we obtained comparability among our data.
4. Inter-University Consortium for Political and Social Research, Study 7534. The consortium distributes a codebook with the data set, but its codebook provides only basic documentation essential to using the data. Information about the data set is available from the consortium at the University of Michigan, Box 1248, Ann Arbor, Michigan 48106, U.S.A.