Wiki surveys

All Our Ideas

Show people two ideas. They pick one. Every vote sharpens a crowd-ranked list of what matters most — and anyone can add ideas of their own.

Run a wiki survey Create in Your Priorities Source on GitHub

Why pairwise?

The simplest honest question

Rating scales get gamed and long surveys get abandoned. A choice between two options takes seconds, works in any language, and is remarkably hard to manipulate — while the mathematics of pairwise comparison turns thousands of quick votes into a robust ranking.

Vote in seconds

Participants compare two ideas at a time — as many or as few pairs as they like. No forms, no friction, anonymous-friendly.

The crowd adds ideas

It’s a survey and a brainstorm at once: anyone can submit a new idea, which immediately starts competing in the ranking.

Rankings you can defend

Every idea gets a score from all votes cast, with live results dashboards — a defensible, transparent picture of group priorities.

From Princeton to Reykjavík

A standalone tool, also built into Your Priorities

All Our Ideas began at Princeton University, led by sociologist Matthew J. Salganik, as a way to collect public input without forcing people into a fixed list of survey options. Early Princeton projects tested the idea; All Our Ideas followed in 2010 as an open-source research platform for “wiki surveys”. An early story from 2010 captured the Princeton origin and student-government use case.

The method was later formalized by Salganik and Karen E. C. Levy in their PLOS ONE paper, Wiki surveys: Open and quantifiable social data collection. It showed how pairwise voting can combine the openness of interviews and brainstorming with the scale, speed and statistical discipline of survey research.

Origin storyRead the Princeton story Paper PDFDownload the wiki surveys paper

Citizens Foundation now carries that lineage forward. We have taken over open-source maintenance, operate a standalone setup at all-our-ideas.citizens.is, and rebuilt wiki surveys as a group type inside Your Priorities, so pairwise ranking can run alongside idea generation, debates and participatory budgeting, with the same AI translation, moderation and analytics underneath.

OpenAI Alignment Assembly

1,000 representative Americans used All Our Ideas to rank and submit concerns about making large language models safe for the public.

Read the story

“AI and You” — New Jersey

More than 2,200 workers voted on pairs of statements about generative AI, producing a rank-ordered list of 96 concerns that drove the AI Task Force’s recommendations to the Governor.

Read the story

Tutorial