42
7. For newer democracies, only governments formed during democratic periods are included
(most central-European countries became functioning democracies in the early 1990s).
8. As reported in table II, my regression analysis found that committees were no stronger in
countries with minority governments than in countries with majority governments.
9. Carey and Shugart considered the possibility of including district magnitude (DM) as a
fourth aspect of electoral design but suggest excluding district magnitude from their rank
ordering due to the large number of possible values associated with DM and because the
effect of DM on the degree of particularism varies depending on the value of the variable
ballot. For these reasons I do not include district magnitude as an element in my measure of
electoral system particularism.
10. I also consider the possibility that committee strength is determined by whether or not the
government is unified or divided. I find no statistically significant difference in committee
strength between unified (majority) and divided (minority) forms of government in
parliamentary democracies.
11. Lijphart standardizes each of the ten variables to summaries the degree of consensus
(Lijphart 1999: 247fn1). In adding Eastern European countries to his dataset I re-standardized
each of the ten variables for all 31 countries included in this study.
12. I assume that the rate of coalition government and the current type of government are
somewhat correlated and for this reason I run separate regressions for each.
13. The results were checked for problems associated with multiple regression analysis. No
problems of multicollinearity or heteroscedasticity arise.
14. Probably the closest proxy would be the electoral system and the degree to which political
parties control access to the ballot. This same variable acts as a measure of particularism.
Convention