But under the first procedure, the party would receive a legislative strength score of .11 and an instability score of 1.5. These scores, calculated under the more straightforward method of using the total number of years in the time period rather than the number of years the party of the legislature operated, seem far more suited to the measurement of' governmental status, for they consider the amount of time that the party was not represented in any legislative body as well as the amount of time it was.
As with BV105 (legislative instability), these calculations are made by the computer from data that the coders have recorded (see Chapter 13) for the proportions of seats that the party held in the lower house of the legislature for each year from 1950 through 1962.
TABLE 4.5a: Early 1960s: Legislative Strength, Recoded
Coding Results: The average party in our study held approximately 30 percent of the seats in the legislature during the second part of our period and somewhat less in the first part. The distribution of these scores, which ran from 0 to 1.0, are given in grouped form in Tables 4.5a and 4.5b. They are decidedly J-shaped in reverse, with the heavy clustering of parties toward the lower end of the distribution being offset by a few parties which held all or virtually all of the seats. The AC codes are moderately high, and there is no significant correlation between BV205 and AC105.
The extent of party following within the electorate usually takes second place to "legislative strength" as an indicator of governmental status because of' even greater problems of cross-national comparability. According to Banks and Textor, about half the countries in 1963 features a competitive or partially competitive electoral system, about a third were considered noncompetitive, and the remainder were unascertainable or ambiguous. They also report that "freedom of group opposition" was present without serious qualification in somewhat less than half of the countries (1963, pp. 86-87). Given this variety in electoral practice and conditions, one would like to have a measure of "popular support" that was not dependent on the electoral system to register an accurate reading. In principle, this might be obtained through survey research, but the political barriers that interfere with free elections and the reporting of accurate results usually obstruct sample surveys completely. While a party's strength at the polls may be a distortion of its support among the people, it nevertheless is a meaningful indicator of status within the government and is adopted as such despite its problems of comparability across nations.
Because we are interested in measuring the basic electoral strength of "the party" as an abstract entity, we try to assess this strength apart from the personality of prominent party candidates. Votes cast for the party's legislative candidates in national elections were preferred to votes cast for the party's candidate for president. Choosing the votes cast for legislative offices has the additional advantage of facilitating comparisons between countries with presidential and parliamentary systems, which do not usually feature a nationwide vote for a party candidate. Moreover, it has some advantage in dealing with coalition presidential candidates common in some countries, although it occasionally presents spe-