Empirical comparisons of voting procedures

Abstract

This article considers decision making in living systems at the levels of the organization, society, or supranational system. It examines voting data from 48 elections over a four-year period that are used to compare different social choice and ranking procedures. Each election had four to six nominees and approximately 1,000 voters. The data indicate that most elections had a majority candidate and that this candidate did not have the most first-place votes about 10 to 20% of the time. Rankings of nominees were compared for the plurality, Hare, and Australian preference procedures. With five nominees, the plurality and Hare rankings differed about 28% of the time, the plurality and Australian preference rankings differed about 55% of the time, and the Hare and Australian preference rankings differed about 62% of the time.Although the data came from a very different source than data previously analyzed by Coombs, Cohen, and Chamberlin (1984). the present conclusions agree closely with theirs in areas where comparisons can be made.

Choosing the Best Social Order: New Principles of Justice and Normative Dimensions of Choice

“…Some 40 years later, such empirical research is still scarce. Studies comparing results of theoretically derived rules with procedures used in practice have been rare (see Fishburn 1986); so have been studies comparing acceptability of various normative principles. Notable experimental work has been done in a related area of distributive justice.…”