THE CONCEPTUAL FRAMEWORK of the ICPP project has no place for "electoral data" as a concept in the analysis of political parties. Instead, electoral data have a place in the parties project for their use in operationalizing indicators of other concepts, namely, "institutionalization" and "governmental status." It proved convenient in data collection, nevertheless, to treat electoral data as a distinct variable cluster. The electoral data were recorded and key punched separately from the other variables and then entered into the computer to calculate values for four variables.

Our electoral data correspond to three primary types of information: (1) percentages of seats held by parties in the lower house of the national legislature, (2) percentages of votes won by the parties in nationwide elections for the lower house of the national legislature, and (3) percentages of votes won by the parties' candidates for president. Information about the percentages of seats held in the legislature is used in scoring two variables: 1.05, "legislative instability," and 2.05, "legislative strength." Information about votes won in election is used in scoring two others: 1.06, "electoral instability," and 2.06, "electoral strength." For most countries, the results of legislative elections are used in scoring the electoral variables, but, for some countries, presidential elections provide better data.

The electoral data used in the ICPP project are limited to our time periods of study for each country, which are stated in Chapter 2. For many countries, party politics before or after our time period differed greatly from party politics during it. Although originally inclined to collect and report electoral data for a longer span of time than included in the project's general focus of 1950 to 1962, we found the collection of accurate data, especially vote returns, outside our time frame to be far more difficult than anticipated. Therefore, we opted to report in tabular form only the data pertaining to our time period. As described below, however, party representation in legislatures through 1978 is reported graphically in Part Two.

The coding sheet used to record our electoral data provided for entering percentages of seats or votes on a yearly basis. The percentage of votes won by a party for a given year in either legislative or presidential elections, depending on which were selected for the country, was entered in the appropriate field on the form for keypunching--alongside the corresponding adequacy-confidence code (AC) indicating the quality of the vote data. The fields for nonelection years were left blank. Thus, a completely blank field for a given year indicates that there was no election that year. But, if there was an election and the party did not exist at the time (either was not yet formed or had terminated), then the percentage field and its AC code were left blank. If the party did exist during an election but could not participate in the election or did not choose to do so, the AC code was made "1." If it participated in a coalition with other parties so that its own strength could not be tallied or if the elections were fraudulent or if data were simply not available, then the AC code became "2." In effect AC codes of 2 or less indicate that the party's electoral strength for that election cannot be estimated with the available data. AC codes of 3 or more were assigned in accordance with the coding definitions in Chapter 2.

The practice in coding legislative seat data differs slightly from that described for election data. Elections are discrete events, sounding out a party's popular strength at separate points in time. A party's legislative strength, however, extends continuously between general elections--subject to vagaries introduced from by-elections, party splits and mergers, and defections of individual members from party to party. Therefore, we record the percentages of seats held by the parties in each year during our time period, not just in the year associated with the election of party members to the legislature. Moreover, while electoral irregularities may invalidate electoral results as indicators of popular strength, rigged elections nevertheless do produce legislative bodies that have a reality of their own--regardless of whether the composition of the chamber reflects true popular support. Thus failure to win legislative seats, for whatever reason, results in zero legislative strength. Similarly, losing one's seats through a coup and dissolution of the legislature also results in zero legislative strength--even though there is no legislature in existence.

go to page 134