|
|||||||||||
not also be highly institutionalized. Therefore, we establish stability in legislative representation as a component of institutionalization independently of the strength of the party's representation. For practical considerations in operationalizing the variable, moreover, we conceptualize the variable in terms of fluctuations in the proportion of seats held over time and refer to it as instability rather than stability. Legislative instability is interpreted in terms of the strength of the party's representation during the entire time period studied for the country in which the party operates. As stated in Chapter 2, this is normally thirteen years but may vary somewhat depending on the country. Fluctuations in legislative representation on a yearly basis across the entire time period are taken to indicate a lack of institutionalization, with the magnitude of the fluctuations judged against the average level of representation. In this sense, strength of representation does enter the conceptualization, for we want to give more weight to a deviation of 10 percent in representation when the mean percentage of seats held by the party is 20 percent than when the mean percentage of seats is 60 percent. (This is handled in the operationalization by dividing the mean deviation by the mean.) Parties that gain legislative seats after the beginning of our time period--even if their strength remains perfectly stable thereafter-cannot be regarded as highly institutionalized throughout our period of interest. Therefore a party's legislative instability score is based on the total period and not just for the years in which the party held seats or even those when it operated. If the party either boycotted legislative elections or was banned from them throughout, however, then its legislative instability score is undefined. Operational Definition. Legislative instability is expressed by the coefficient of relative variation in percentage of seats (X) held in the lower house of the legislature over the number of years (N) in the time period for its country, as given by the formula
Coding Results. The coders did not apply this formula themselves to the data on legislative representation. Their task was limited to recording the proportion of seats that a party held in the lower house of the legislature during each year of our time period. These data were then processed by a computer programmed to calculate BV 105. If the party held legislative seats at any time during our period of interest, the computer calculated its "legislative instability" score. These scores, which had computed values from 0 to 1.82, are grouped into categories in Table 3.7 to aid comprehension. As the "missing" data entry reveals, we were able to score about 95 percent of the parties on legislative instability, excluding only those parties that never held any seats because they were outlawed during their existence throughout our period. There is a significant relationship between low data quality and high instability as shown by the correlation of -.34 between BV105 and AC105. This is not surprising, for it was often difficult to obtain information about the proportions of legislative seats held by parties over time in countries with emerging party systems. The mean score for AC105 is |
|||||||||||