The Voting Procedure Advisor app below helps users identify which voting rule best fits their decision situation. It implements a decision tree based on widely studied properties of voting procedures — such as majority guarantees, susceptibility to strategic voting, and proportionality.

How the app works?

The app implements a decision tree. The decision tree proceeds in these steps:

1. Define the structure of the decision (number of voters, single or multiple winners).
2. Identify desired procedural properties (simplicity, manipulation resistance, auditability).
3. Evaluate ballot design constraints (ranking, scoring, or approval).
4. Apply secondary filters (speed, proportionality, majority requirement).
5. Output a ranked list of methods that best fit the stated priorities.

Each question narrows the field of possible voting systems by adjusting preference weights that reflect theoretical and practical trade-offs.

1. Group size

Question: How many people will vote?
Purpose: Adjusts later priorities for feasibility, complexity, and transparency.

• 1–5: Small groups can use deliberative, expressive methods (e.g., Condorcet or Borda).
• 6–20: Mixed methods like IRV or Approval are suitable — manageable yet representative.
• 21–100: Medium groups need scalable, moderately complex rules (IRV, Approval, STV).
• 100+: Very large groups need simple, auditable methods (Plurality, Approval, Party-List).

This parameter doesn’t directly select a method but influences others (simplicity, auditability).

2. Winners

Question: Are you selecting one winner or multiple winners?
Branching: This is the first major split in the tree.

• Single-winner branch: leads to rules such as Plurality, Runoff, IRV, Condorcet, Approval, Score, or Borda.
• Multi-winner branch: leads to STV, SNTV, or Party-List systems.

The rest of the tree is evaluated within the relevant branch only.

3. Simplicity

Question: How important is simplicity?
Purpose: Balances usability versus representational accuracy.

• Very important: Emphasizes transparent, fast, easy-to-audit methods (Plurality, Approval, SNTV).
• Somewhat important: Accepts moderate complexity for accuracy (IRV, Runoff).
• Not important: Allows sophisticated tabulations for fairness (Condorcet, STV).

This score interacts strongly with group size and auditability.

4. Manipulation resistance

Question: How worried are you about strategic voting?
Purpose: Identifies tolerance for tactical behavior.

• High concern: Favors Condorcet and IRV — less manipulable than Plurality or Borda.
• Medium concern: Prefers moderate safeguards such as Runoff or Approval.
• Low concern: Simpler systems are acceptable even if manipulable.

This question differentiates Condorcet from Borda and Plurality within the single-winner branch.

5. Intensity of preference

Question: Should the method capture strength of preference?
Purpose: Determines whether voters should express how strongly they support options.

• Yes: Points to Approval, Score (Range Voting), or Borda Count — all measure intensity.
• No: Focus remains on order or binary choice (IRV, Condorcet, Plurality, Runoff).

This dimension introduces score-based methods that otherwise would be excluded.

6. Majority guarantee

Question: Do you want a guaranteed majority-supported winner when one exists?
Purpose: Enforces the majoritarian principle.

• Yes: Increases weight of Runoff, IRV, Condorcet — which ensure majority support or near-majority outcomes.
• No: Allows methods that may pick broadly supported but not majority candidates (Approval, Borda, Score).

This parameter refines the single-winner path between expressiveness and legitimacy.

7. Proportionality (multi-winner only)

Question: For multi-winner outcomes, how important is proportional representation?
Purpose: Distinguishes between proportional and majoritarian multi-seat outcomes.

• Very important: STV or Party-List methods.
• Somewhat important: Still favors STV or Party-List, with tolerance for minor deviations.
• Not important: Simpler rules like SNTV are sufficient.

This node is active only if the user selected multiple winners earlier.

8. Auditability

Question: How important is auditability (transparent counting, easy recounts)?
Purpose: Ensures trust in large or contested elections.

• Very important: Prioritizes methods with simple arithmetic (Plurality, Approval, Runoff, Party-List, SNTV).
• Somewhat important: Allows moderate complexity (IRV, Condorcet).
• Not important: Does not constrain the method.

This influences large-scale decisions where legitimacy depends on public verifiability.

9. Ballot complexity tolerance

Question: What ballot complexity will voters tolerate?
Purpose: Filters out infeasible ballot formats.

• Pick one only: Limits methods to Plurality, Runoff, SNTV.
• Rank candidates: Enables IRV, Condorcet, Borda, STV.
• Approve or score: Allows Approval, Score, and hybrid Party-List variants.

This question structurally determines which family of methods remains viable.

10. Rounds and decision speed

Question: Is minimizing the number of rounds critical?
Purpose: Weighs efficiency against deliberation.

• Yes: Prioritizes one-round methods (Plurality, Approval, Score, IRV).
• No: Allows multi-round methods (Two-Round Runoff).

This final constraint filters out slower but sometimes fairer systems.

11. Integration logic

Each answer contributes weighted scores to potential methods.
For instance:

• “Single winner + Very simple + Highly worried about manipulation” → IRV or Condorcet.
• “Large group + Multi-winner + Very proportional + One round” → Party-List or STV.
• “Small group + Intensity yes + No majority requirement” → Score or Borda.
• “One winner + Very simple + One round” → Plurality or Approval.
• “Multi-winner + Not proportional + Simple ballots” → SNTV.*

The algorithm ranks all methods by cumulative score and returns:

1. The recommended method (highest score).
2. Two alternative options (next highest).
3. A list of reasons explaining the fit between answers and method properties.

12. Recommendation output

The app’s output includes:

• Top three methods, ordered by computed suitability.
• Qualitative rationale, generated from the decision path (e.g., “You emphasized simplicity; this method is easy to explain”).
• Metadata: timestamp and user responses.
• Optional JSON export for documentation or comparison.

Summary

Dimension	Purpose	Typical Influence
Group size	Feasibility	Favors simple or complex methods
Winners	Structural branch	Single vs multi-winner systems
Simplicity	Ease vs accuracy	Penalizes complex methods
Manipulation concern	Strategic robustness	Rewards Condorcet / IRV
Intensity	Expressiveness	Rewards Score / Approval
Majority	Legitimacy	Rewards IRV / Runoff
Proportionality	Representation	Rewards STV / Party-List
Auditability	Trust & transparency	Rewards Plurality / Approval
Ballot complexity	Voter capacity	Filters families of methods
Rounds	Efficiency	Rewards one-round methods

In short:

The decision tree implemented in the app maps decision context to procedural design. It helps users choose a voting method that balances fairness, simplicity, and practicality for their specific situation — ensuring that the rule chosen reflects both the structure of the group and its governance priorities.