pith. sign in

arxiv: 1906.10562 · v1 · pith:5B7H5IPUnew · submitted 2019-06-25 · 💻 cs.AI

Awareness of Voter Passion Greatly Improves the Distortion of Metric Social Choice

Ben Abramowitz , Elliot Anshelevich , Wennan Zhu This is my paper

Pith reviewed 2026-05-25 16:38 UTC · model grok-4.3

classification 💻 cs.AI
keywords metric social choicedistortionvoting mechanismspreference strengthssocial costordinal preferences
0
0 comments X

The pith

Knowing a small amount of information about how strongly voters prefer candidates greatly reduces distortion in metric voting.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops voting rules for settings where voters and candidates sit in an unknown metric space and the goal is to minimize total distance from voters to the winner. It shows that mechanisms given even a modest amount of data on preference intensities achieve substantially lower distortion than rules that see only rankings. The work measures how distortion improves as the amount or type of strength information grows and identifies which forms of extra data help most. A sympathetic reader would care because real elections often reveal whether a voter feels strongly or weakly about a candidate, and using that signal could produce outcomes closer to the true optimum without requiring full numerical utilities.

Core claim

In an arbitrary metric space, mechanisms that receive a small amount of information about voter preference strengths select candidates whose social cost is a much smaller multiple of the optimal candidate's cost than is possible with ordinal information alone; the paper quantifies the resulting distortion trade-offs and also defines an ideal-candidate distortion that compares the chosen winner against the best conceivable candidate rather than only those in the election.

What carries the argument

Voting mechanisms that incorporate limited preference-strength data to choose a candidate minimizing total voter distance in an unknown metric.

If this is right

  • Distortion bounds drop sharply once even minimal strength data is available.
  • The improvement can be traded off against the quantity or precision of the extra information supplied.
  • Different kinds of strength information produce measurably different improvements.
  • The gap between the chosen candidate and the single best possible candidate (ideal distortion) can be bounded separately from ordinary distortion.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Eliciting a few intensity comparisons may be a low-cost way to improve real-world voting outcomes without full cardinal utilities.
  • The same strength signals could be tested in related metric problems such as facility location or clustering.
  • If strength data can be obtained cheaply, the paper's trade-off curves give a practical way to decide how much to collect.

Load-bearing premise

Some limited information about the strengths of voter preferences is known in addition to their rankings.

What would settle it

A concrete metric space and set of voter preferences in which every mechanism that uses the stated strength information still produces distortion at least as high as the best ordinal-only mechanism.

Figures

Figures reproduced from arXiv: 1906.10562 by Ben Abramowitz, Elliot Anshelevich, Wennan Zhu.

Figure 1
Figure 1. Figure 1: Distortion for two candidates with pref [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 3
Figure 3. Figure 3: Best achievable distortion for a single threshold [PITH_FULL_IMAGE:figures/full_fig_p018_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: Best known distortion for multiple can￾didates if allowed the best choice of m thresholds. Converges to 2 with the number of thresholds. 22 [PITH_FULL_IMAGE:figures/full_fig_p022_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Tradeoff between δI and ∆ with voter preferences and a single threshold τ in the two can￾didates setting [PITH_FULL_IMAGE:figures/full_fig_p027_6.png] view at source ↗
read the original abstract

We develop new voting mechanisms for the case when voters and candidates are located in an arbitrary unknown metric space, and the goal is to choose a candidate minimizing social cost: the total distance from the voters to this candidate. Previous work has often assumed that only ordinal preferences of the voters are known (instead of their true costs), and focused on minimizing distortion: the quality of the chosen candidate as compared with the best possible candidate. In this paper, we instead assume that a (very small) amount of information is known about the voter preference strengths, not just about their ordinal preferences. We provide mechanisms with much better distortion when this extra information is known as compared to mechanisms which use only ordinal information. We quantify tradeoffs between the amount of information known about preference strengths and the achievable distortion. We further provide advice about which type of information about preference strengths seems to be the most useful. Finally, we conclude by quantifying the ideal candidate distortion, which compares the quality of the chosen outcome with the best possible candidate that could ever exist, instead of only the best candidate that is actually in the running.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 3 minor

Summary. The paper develops voting mechanisms for metric social choice that minimize social cost (total distance to the chosen candidate) when voters and candidates lie in an arbitrary unknown metric. It assumes access to a small amount of information about preference strengths in addition to ordinal rankings, constructs mechanisms that achieve substantially lower distortion than ordinal-only mechanisms, quantifies the resulting tradeoffs as a function of the amount and type of strength information, offers guidance on the most useful forms of such information, and analyzes the distortion relative to an ideal candidate that may not be in the candidate set.

Significance. If the mechanisms and distortion bounds are correct, the work is significant because it demonstrates that even minimal cardinal information can produce large, quantifiable improvements over the well-studied ordinal-distortion setting, while remaining within the metric social-choice framework. The explicit tradeoff curves and the ideal-candidate analysis provide concrete guidance that is absent from prior ordinal-only results.

major comments (2)
  1. [§4, Theorem 4.3] §4, Theorem 4.3: the claimed distortion bound of 3 appears to rely on the specific form of the 'passion' information (a single bit per voter indicating whether the top choice is at least twice as good as the second); the proof sketch does not explicitly verify that this bound remains valid for arbitrary unknown metrics when the bit is adversarially placed.
  2. [§5.2] §5.2, the tradeoff quantification: the paper states that distortion improves continuously with the number of strength bits, yet the plotted curves in Figure 3 are obtained only for a discrete set of bit budgets; it is unclear whether the continuous interpolation is supported by a matching lower-bound construction or is merely an upper-bound interpolation.
minor comments (3)
  1. [§2] Notation: the symbol d(v,c) is used both for the true metric distance and for the reported cost; a brief clarification in §2 would avoid confusion.
  2. [Figure 2] Figure 2: the y-axis label 'distortion' should specify whether it is worst-case or average-case over the metric.
  3. [§1] Related work: the discussion of Anshelevich et al. (2018) on ordinal distortion omits the subsequent improvement to distortion 3 by Gkatzelis et al.; adding this reference would strengthen the comparison.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful review and constructive comments. We address each major comment below.

read point-by-point responses

  1. Referee: [§4, Theorem 4.3] the claimed distortion bound of 3 appears to rely on the specific form of the 'passion' information (a single bit per voter indicating whether the top choice is at least twice as good as the second); the proof sketch does not explicitly verify that this bound remains valid for arbitrary unknown metrics when the bit is adversarially placed.

    Authors: The analysis in Theorem 4.3 is intended to apply to arbitrary unknown metrics, with the passion bit chosen adversarially. The proof bounds the social cost using only the triangle inequality and the information consistent with the ordinal rankings plus the passion bit. We agree the sketch could state this generality more explicitly and will add a short clarifying sentence. revision: partial

  2. Referee: [§5.2] the tradeoff quantification: the paper states that distortion improves continuously with the number of strength bits, yet the plotted curves in Figure 3 are obtained only for a discrete set of bit budgets; it is unclear whether the continuous interpolation is supported by a matching lower-bound construction or is merely an upper-bound interpolation.

    Authors: The curves in Figure 3 report the distortion achieved by our mechanisms at discrete bit budgets. The claim of continuous improvement follows because additional bits cannot worsen the bound. The plotted lines interpolate the discrete points for readability and do not assert a matching continuous lower bound. We will revise the text and caption to clarify this. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation relies on new mechanism constructions independent of inputs.

full rationale

The paper develops novel voting mechanisms that incorporate limited preference-strength information in unknown metric spaces to achieve improved distortion bounds over ordinal mechanisms, while quantifying tradeoffs and ideal-candidate distortion. No self-definitional reductions, fitted inputs renamed as predictions, or load-bearing self-citation chains appear in the provided abstract or description. The central claims rest on explicit mechanism constructions and distortion analyses that are externally benchmarked against prior ordinal distortion results in social choice, making the derivation self-contained.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The paper relies on standard metric space assumptions and the availability of preference strength data, with no free parameters or invented entities mentioned in the abstract.

axioms (2)
  • domain assumption Voters and candidates are embedded in an arbitrary unknown metric space where distance represents cost.
    Stated in the abstract as the setting for the problem.
  • domain assumption A small amount of information about preference strengths is available in addition to ordinal preferences.
    The key assumption enabling the new mechanisms.

pith-pipeline@v0.9.0 · 5723 in / 1235 out tokens · 38668 ms · 2026-05-25T16:38:39.300765+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

31 extracted references · 31 canonical work pages · 1 internal anchor

  1. [1]

    Approximating optimal social choice under metric preferences

    Elliot Anshelevich, Onkar Bhardwaj, Edith Elkind, John Postl, and Piotr Skowron. Approximating optimal social choice under metric preferences. Artificial Intelligence, 264:27–51, 2018

    work page 2018

  2. [2]

    Randomized social choice functions under metric preferences

    Elliot Anshelevich and John Postl. Randomized social choice functions under metric preferences. Journal of Artificial Intelligence Research (JAIR), 58:797–827, 2017

    work page 2017

  3. [3]

    Ordinal Approximation for Social Choice, Matching, and Facility Location Problems Given Candidate Positions

    Elliot Anshelevich and Wennan Zhu. Ordinal Approximation for Social Choice, Matching, and Facility Location Problems Given Candidate Positions. In International Conference on Web and Internet Economics (WINE), pages 3–20. Springer, 2018. 32

    work page 2018

  4. [4]

    Advances in the spatial theory of voting

    Kenneth Arrow. Advances in the spatial theory of voting. Cambridge University Press, 1990

    work page 1990

  5. [5]

    Preference elicitation for participatory budgeting

    Gerdus Benade, Swaprava Nath, Ariel D Procaccia, and Nisarg Shah. Preference elicitation for participatory budgeting. In Thirty-First AAAI Conference on Artificial Intelligence (AAAI) , pages 376–382, 2017

    work page 2017

  6. [6]

    Low-Distortion Social Welfare Functions

    Gerdus Benade, Ariel D Procaccia, and Mingda Qiao. Low-Distortion Social Welfare Functions. In Thirty-Third AAAI Conference on Artificial Intelligence (AAAI), 2019

    work page 2019

  7. [7]

    Truthful and near-optimal mechanisms for welfare maximization in multi-winner elections

    Umang Bhaskar, Varsha Dani, and Abheek Ghosh. Truthful and near-optimal mechanisms for welfare maximization in multi-winner elections. In Thirty-Second AAAI Conference on Artificial Intelligence (AAAI), pages 925–932, 2018

    work page 2018

  8. [8]

    Primarily about Primaries

    Allan Borodin, Omer Lev, Nisarg Shah, and Tyrone Strangway. Primarily about Primaries. In Thirty-Third AAAI Conference on Artificial Intelligence (AAAI), 2019

    work page 2019

  9. [9]

    Optimal social choice functions: A utilitarian view

    C Boutilier, I Caragiannis, S Haber, T Lu, A D Procaccia, and O Sheffet. Optimal social choice functions: A utilitarian view. Artificial Intelligence, 227:190–213, 2015

    work page 2015

  10. [10]

    Social choice and intensity of preference

    Donald E Campbell. Social choice and intensity of preference. Journal of Political Economy , 81(1):211–218, 1973

    work page 1973

  11. [11]

    Subset Selection Via Implicit Utilitarian V oting

    I Caragiannis, S Nath, A D Procaccia, and N Shah. Subset Selection Via Implicit Utilitarian V oting. Journal of Artificial Intelligence Research (JAIR), 58:123–152, 2017

    work page 2017

  12. [12]

    V oting almost maximizes social welfare despite limited communication

    Ioannis Caragiannis and Ariel D Procaccia. V oting almost maximizes social welfare despite limited communication. Artificial Intelligence, 175(9-10):1655–1671, 2011

    work page 2011

  13. [13]

    Of the people: voting is more effective with repre- sentative candidates

    Yu Cheng, Shaddin Dughmi, and David Kempe. Of the people: voting is more effective with repre- sentative candidates. In Proceedings of the 2017 ACM Conference on Economics and Computation (EC), pages 305–322. ACM, 2017

    work page 2017

  14. [14]

    On the distortion of voting with multiple repre- sentative candidates

    Yu Cheng, Shaddin Dughmi, and David Kempe. On the distortion of voting with multiple repre- sentative candidates. In Thirty-Second AAAI Conference on Artificial Intelligence (AAAI) , pages 973–980, 2018

    work page 2018

  15. [15]

    Random Dictators with a Ran- dom Referee: Constant Sample Complexity Mechanisms for Social Choice

    Brandon Fain, Ashish Goel, Kamesh Munagala, and Nina Prabhu. Random Dictators with a Ran- dom Referee: Constant Sample Complexity Mechanisms for Social Choice. Thirty-Third AAAI Conference on Artificial Intelligence (AAAI), 2019

    work page 2019

  16. [16]

    Preference intensity measurement

    Peter H Farquhar and L Robin Keller. Preference intensity measurement. Annals of operations research, 19(1):205–217, 1989

    work page 1989

  17. [17]

    On voting and facility location

    Michal Feldman, Amos Fiat, and Iddan Golomb. On voting and facility location. In Proceedings of the 2016 ACM Conference on Economics and Computation (EC), pages 269–286. ACM, 2016

    work page 2016

  18. [18]

    Preference intensity representation and revelation

    Georgios Gerasimou. Preference intensity representation and revelation. School of Economics and Finance Discussion Paper No. 1716. 2019

    work page 2019

  19. [19]

    On the Distortion Value of the Elections with Abstention

    Mohammad Ghodsi, Mohamad Latifian, and Masoud Seddighin. On the Distortion Value of the Elections with Abstention. Thirty-Third AAAI Conference on Artificial Intelligence (AAAI) AAAI Conference on Artificial Intelligence (AAAI), 2019

    work page 2019

  20. [20]

    Metric distortion of social choice rules: Lower bounds and fairness properties

    Ashish Goel, Anilesh K Krishnaswamy, and Kamesh Munagala. Metric distortion of social choice rules: Lower bounds and fairness properties. In Proceedings of the 2017 ACM Conference on Economics and Computation (EC), pages 287–304. ACM, 2017

    work page 2017

  21. [21]

    A Unified Theory of Voting: Directional and Proximity Spatial Models

    Bernard Grofman and Samuel Merrill III. A Unified Theory of Voting: Directional and Proximity Spatial Models. Cambridge University Press, 1999

    work page 1999

  22. [22]

    V ote until two of you agree: Mechanisms with small distortion and sample complexity

    Stephen Gross, Elliot Anshelevich, and Lirong Xia. V ote until two of you agree: Mechanisms with small distortion and sample complexity. In Thirty-First AAAI Conference on Artificial Intelligence (AAAI), 2017

    work page 2017

  23. [23]

    The spatial theory of voting: an introduction

    Melvin J Hinich and James M Enelow. The spatial theory of voting: an introduction . Cambridge University Press Cambridge,, UK, 1984

    work page 1984

  24. [24]

    H. Moulin. Choosing from a tournament. Social Choice and Welfare, 3(4):271–291, 1986. 33

    work page 1986

  25. [25]

    Improved Metric Distortion for Deterministic Social Choice Rules

    Kamesh Munagala and Kangning Wang. Improved Metric Distortion for Deterministic Social Choice Rules. Proceedings of the 2019 ACM Conference on Economics and Computation (EC) , 2019

    work page 2019

  26. [26]

    A decade of experimental research on spatial models of elections and commit- tees

    C Ordeshook Peter. A decade of experimental research on spatial models of elections and commit- tees. Advances in the spatial theory of voting, page 99, 1990

    work page 1990

  27. [27]

    Approval-Based Elections and Distortion of Voting Rules

    Grzegorz Pierczy ´nski and Piotr Skowron. Approval-Based Elections and Distortion of V oting Rules. arXiv preprint arXiv:1901.06709, 2019

    work page internal anchor Pith review Pith/arXiv arXiv 1901

  28. [28]

    Procaccia and Jeffrey S

    Ariel D. Procaccia and Jeffrey S. Rosenschein. The Distortion of Cardinal Preferences in V oting. In 10th International Workshop on Cooperative Information Agents (CIA), pages 317–331. Springer, 2006

    work page 2006

  29. [29]

    The spatial model of politics

    Norman Schofield. The spatial model of politics. Routledge, 2007

    work page 2007

  30. [30]

    Social choice under metric preferences: scoring rules and STV

    Piotr Krzysztof Skowron and Edith Elkind. Social choice under metric preferences: scoring rules and STV. In Thirty-First AAAI Conference on Artificial Intelligence (AAAI), pages 706–712, 2017

    work page 2017

  31. [31]

    intensity

    Kendall Willmoore and George W. Carey. The “intensity” problem and democratic theory. Ameri- can Political Science Review, 62(1):5–24, 1968. 34

    work page 1968