ARTFEED — Contemporary Art Intelligence

New Indices Measure Agreement, Diversity, and Polarization in Approval Elections

other · 2026-05-16

A new computer science paper proposes indices to quantify agreement, diversity, and polarization in approval elections, where voters approve or disapprove of candidates. The indices output values between 0 and 1 and are normalized for saturation, meaning elections with different average approval rates but similar structures yield comparable values. The authors derive a new map of approval elections and analyze real-life data from Pabulib, Preflib, and other sources to highlight similarities and differences. The work is published on arXiv under computer science and game theory.

Key facts

  • Indices output values between 0 and 1 for agreement, diversity, and polarization.
  • Indices are normalized with respect to saturation.
  • Applied to real-life elections from Pabulib, Preflib, and other sources.
  • Paper is categorized under Computer Science > Computer Science and Game Theory.
  • Published on arXiv with ID 2605.14983.
  • Authors propose several indices and analyze their properties.
  • A new map of approval elections is derived.
  • The work shows similarities and differences between various real-life elections.

Entities

Institutions

  • arXiv
  • Pabulib
  • Preflib

Sources