A Different World:
Relative Causal Inference and the Study of Mixed Electoral Systems
“My friend, the oracle says, to you I shall speak the simple truth,
which is simply that the cat-dog does not exist.” (Sartori, 1991: 247)
1. The Problem
Over twenty years ago, Riker (1982) chose to chronicle the formulation, the attempts at empirical
falsification, the repeated revisions, and the eventual acceptance of Duverger’s law and
hypothesis
1
by a majority of scholars in the field as an example of the fact that political science,
like every other truly scientific discipline, has a history characterized by the accumulation of
knowledge. Riker’s (1982) choice was hardly coincidental. Voting and, more generally, the
consequences that electoral rules have on the strategic behavior of relevant political actors are
among the most widely investigated sets of research problems in political science. The literature
on the outcomes produced by electoral institutions, whether descriptive or analytical, formal or
empirical, single-country or cross-national, is, in terms of accumulation of knowledge, perhaps
the best developed of any subfield within the discipline.
Not unlike Duverger’s propositions, much of the scholarly work on the topic has attempted to
comprehend the heterogeneity in outcomes generated by the profoundly diverse families of
electoral systems that are in use worldwide. The fundamental distinction between the most
widely adopted types of electoral rules lies, in rather broad terms, in the fact that some systems
elect candidates through contests held in single-member districts (SMD, which may allocate
seats according to either the majority or plurality formulae
2
) while others, commonly referred to
as proportional representation (PR), allocate seats to parties in multi-member districts in rough
proportion to their vote share.
3
Extensive evaluations have been conducted on, among other
things, how alternative electoral systems affect a voter’s propensity to vote sincerely or
sophisticatedly, how strategic voting gives rise to district-level party systems,
4
how electoral
Convention