Picking Fair Committees Just Got Faster: New Math Shows How

Imagine your neighborhood is choosing a committee to plan events. Everyone votes, but no committee can make everyone happy. The goal is a group that proportionally reflects voters' interests. Now imagine an algorithm that proposes a committee, gets told it's not fair enough, and tries again. How many tries does it need?

This new math paper answers that question. It looks at two types of feedback: full feedback, which explains exactly which voter group was underrepresented, or candidate-only feedback, which just names a missing person. Surprising news: for most fairness standards, the number of rejected proposals depends only on the size of the committee — not on how many people vote or run. So even if your neighborhood grows from 100 to 100,000 people, the algorithm needs about the same number of tries.

Why does this matter? In real life, things like citizen panels, corporate boards, or even platform moderation teams face similar challenges. Sometimes fairness is about making sure no group is left out. Knowing that a good committee can be found quickly means less time arguing, fewer bureaucratic delays, and lower costs for organizers. It also gives practical guidance for building software that helps groups make fair decisions.

Of course, this is theoretical work. The algorithms are proofs of concept, not ready-made apps. And the fairness standards studied here are specific: proportional justified representation (PJR+) and extended justified representation (EJR+). But for anyone who has ever sat through a long, painful committee-selection meeting, the idea that math can make it faster is genuinely good news.