In recording percentages of seats for keypunching, the AC codes are assigned in accordance with the coding definitions in Chapter 2. A blank AC code is assigned only in years for which the party did not exist; otherwise, we try to score each party for percentage of seats held in the lower house of the legislature in each year during the time period for that country. When the literature fails to provide information as to the distribution of seats in years between elections, we extrapolate the results of the most recent election. Sometimes we are led to adjust our adequacy-confidence ratings in accordance with our understanding of the stability of party politics in the country. For countries with volatile parties, we are apt to reduce our AC codes as we extrapolate the percentage of seats won to years beyond the election year. For countries with more predictable politics, we hold firmer to our original AC codes, although they usually are reduced by one point to indicate extrapolation.

After being recorded and keypunched, the electoral data were processed with a computer program which generated our measures of legislative and electoral instability and legislative and electoral strength (see variables 1.05, 1.06, 2.05, and 2.06). The results are reproduced in Part Two following the introductory essays on party politics. The computation of legislative and electoral instability is discussed in Chapter 3 under basic variables 1.05 and 1.06. See the definitions of basic variables 2.05 and 2.06 in Chapter 4 for discussions of legislative and electoral strength.

Part Two contains a graph for each country that plots the percentage of seats in its lower legislative chamber held by each of its parties from 1950 through 1978. These graphs report not only the proportion of seats held by the 158 parties in the study but also the seats held by 50 additional parties that met our "strength" and "stability" requirements (see Chapter 1) after 1963, which qualified them for study in the 1963-1978 period. These parties and their years of origin are listed in Chapter 15. They are also described briefly in Part Two as "new" parties.

Because the post-1963 period exceeded the scope of the ICPP project, the 1963-1978 legislative seat data were added many years after financial support for the project had terminated. This meant that data on party strength after 1963 were collected differently, with much greater reliance on tertiary sources and less compulsion to track down differences between sources of only one or two percentage points. One consequence of the research modifications was an inability to obtain adequate information on the percentages of votes that parties won in general elections. For about 25 percent of the countries, voting returns for all elections since 1963 simply were not readily available in tertiary or convenient secondary sources and, based on our experience with studying 1950-1962, would have required substantial effort to locate. We opted instead to limit our concern with party strength since 1963 to legislative representation, which could be determined much more easily from the available sources.

Numerous sources were consulted for data on party legislative representation. Among the most useful were the annual publications of Europa Yearbook, Political Handbook of the World, World Strength of Communist Party Organizations, Africa South of the Sahara, and the Yearbook of International Communist Affairs. The biweekly issues of Keesing's Contemporary Archives were invaluable for explaining party splits and electoral alliances. The Army Area Handbook series of "country studies" often furnished information available nowhere else on countries in the Third World. For "Western" countries, the best source of seat data up to 1970 was the International Almanac of Electoral Statistics (Mackie and Rose 1974). For African countries, the best source up to 1970 was Welfing (1971). Many other sources, too numerous to mention, were consulted for legislative seat distributions in particular countries.

Notwithstanding the heavy reliance on tertiary sources, data on party legislative representation were recorded with the same data quality concerns as reflected throughout the course of the ICPP project, with adequacy-confidence codes being assigned to each party's percentage for each year. Adding the 1963-1978 legislative seat data to the seat data collected for 1950 to 1962 produced a set of 29 variables pertaining to legislative strength over time. These data were converted into time-series form and graphed as illustrated in Part Two. The system instability scores that accompany the graphs were calculated as indicated in Chapter 15.

return to Table of Contents