John Gerring, "Minor Parties in Plurality Electoral Systems," Party Politics, 11 (January, 2005), 79-107.

First Paragraph:
To what extent do electoral systems determine party systems? Are plurality electoral systems invariably dominated by two parties? Under what conditions is Duverger's law true? These are central questions in the study of electoral politics, and of critical concern to countries currently employing plurality rules as well as to others who may be contemplating a switch to a more majoritarian system.

Figures and Tables:
Table 1. Countries and key variables
Figure 1. Minor party performance. Mean value for period under study
Table 2. Regression results

First paragraph in the conclusion:
Without deprecating the role of minor party leaders and activists (at the margins, leadership certainly matters) it may be put forth as a general hypothesis that minor party performance is endogenous to major party performance. Minor parties owe their successes or failures more to cracks in the armor of the major parties than to their own efforts (Schattschneider, 1942, 1960). Indeed, over the long haul (ignoring the results of particular elections and electoral periods), minor parties are best considered as a characteristic of a polity, rather than of a particular group or grievance. The level of success that these protest groups achieve over time - the equilibrium result of a party system - is a system-wide phenomenon. Where the major parties are doing their job, rumblings of discontent should be coopted (or if you prefer, represented) by one or both of these established institutions. It is only when they fail to perform this representative function, or when their performance is purely rhetorical (and hence not credible), that minor parties have a good opportunity to win seats in a plurality electoral system.