THE International Comparative Political Parties (ICPP) project was established in 1967 for the purpose of conducting the first comprehensive, empirically based,comparative analysis of political parties throughout the world (Janda 1968b; Janda 1969; Janda 1970c). The project focuses on 158 political parties operating in 53 countries during 1950-1962 and traces their fate through 1978. All but three of these countries were selected at random, with five countries being drawn from each of ten cultural geographical regions of the world.(See Table 1.1 for regions, countries, and distribution of parties.) This stratified random sample, constituting approximately 50 percent of the world's party systems, was later augmented by the addition of the United States, the United Kingdom, and Canada, none of which appeared in the original sample.

The procedures for collecting data on the 158 parties in the project differed greatly for the 150 parties from the original sample of 50 countries as opposed to the 8 parties from the United States, the United Kingdom, and Canada. The American and British parties were coded by native students of politics in both countries employing conventional methods of library research to aid their judgment. The Canadian parties were coded as a project of Northwestern's Canadian Studies Program. Parties in all other countries were coded primarily by nonnative research assistants using a modern microfilm and computer information retrieval system to search our files of relevant literature for information to guide their coding.

The information files that supported their research were created during the bibliographic and indexing phases of the project, which lasted from 1967 through 1969. Intensive bibliographic searches for material on party politics in each of the original 50 countries turned up thousands of citations to material that seemed pertinent. Most of the material was rejected after examination, but more than 3,000 documents were judged to be useful for our research and were processed into our information files. This meant (1) that each relevant page of each document was read and tagged for substantive content with one or more "index codes"--given in abbreviated form in Table 1.2 and (2) that each page was photographed on microfilm with its accompanying index codes also encoded on film in machine-readable binary form to permit rapid search and retrieval (see Janda 1967; Janda 1968c). The distribution of material in the files by country is shown in Table 1.1.

Even aided by this computerized microfilm information retrieval system, the coding phase of the ICPP project lasted from 1969 to 1973. Three reasons may be cited for the delay. (1) Codes had to be assigned to a large number of variables, most of which required finer judgments about scoring decisions than normally are involved in cross-national data collections. (2) There was an unusually heavy investment in data quality procedures, which extended to the tagging of every coding decision with a data quality code and to the description of these coding decisions with a sentence or a paragraph suitable for publication with the raw data. (3) Because only one retrieval machine was available for most of the research period, the coding had to proceed in serial fashion, with parties in each country being coded in turn rather than simultaneously, with many researchers engaged in coding all the parties.

The manual of instructions that guided coding in the ICPP project went through four editions and grew to more than 200 pages in length (Janda 1972). It described the twelve major concepts that constituted the conceptual framework of the ICPP project and organized some 100 "basic" variables into twelve "clusters" of variables that were proposed as indicators of the major concepts. Each basic variable was discussed first in terms of its conceptual basis and then its operational definition. Part One of this volume consists partly of an abridged version of that coding manual, omitting instructions for using the retrieval equipment to search for information pertinent to coding a given variable. In place of these search instructions, Part One presents a discussion of the "coding results," which reports the summary statistics for the variable along with commentary about problems and peculiarities encountered at both the conceptual and operational levels. Part Two presents the codes assigned to individual parties on each

go to page 4