Consensus Dynamics Project

This is a mathematical modeling project, looking at the problem of how a group of people makes a decision together when participants may disagree about what is desirable. This project is about making idealizations of that situation, looking at it as a dynamic process, and investigating via models what strategies for reaching an agreeable solution might be more and less effective.

Here and here are more detailed descriptions of the project.

News from the consensus research project

• I first described this project in 2009 in a a more general research proposal (p. 4).

• I wrote down the basic model structure, with a group of people searching together in a space of possible solutions to a problem, in January 2010.

• I created these public wiki pages and did the first simulation experiments in June 2011:
  • A baseline experiment, confirming that a "group" of only one person is able to come up with a proposal that's better for that person than the obvious alternatives.
  • A nontrivial experiment in which people offer "friendly amendments" to one other's proposals, until either no one has blocking concerns or no one can think of anything to add. These experiments assume very rugged "fitness landscapes" (no similarity between closely related proposals) and very diverse groups (different people's preferences have nothing in common), and a rudimentary process in which people are opaque about how they arrive at their proposals. Here larger groups get stuck, because more people means less chance of everyone agreeing. More degrees of freedom seems to help.

• In July 2011 I expanded the simulation model to look at the effects of independently varying "facilitation strategy", "proposal strategy", and "block strategy". The results of this experiment suggest that the "facilitation strategies" I considered don't make as much difference as the other two strategies, and that again more degrees of freedom help and larger group size hurts.
  • I presented these early results among other things in a public talk on July 15, 2011.
  • I also created the literature review section at the same time.

• This project was the subject of a #SciFund Challenge crowdfunding appeal on rockethub.com, which ran Nov. 1 - Dec. 15, 2011 and raised about $1300. I used that funding to support a month of focused research in March 2012.

• Nov. 21, 2011: Notes from Greece.

• Nov. 27, 2011: Notes on distributed computing: is it useful to draw an analogy between distributed computing and working in a leaderless group?

• Dec. 16, 2011: I present this project at McMaster University.

• Dec. 22, 2011: The Pizza Exercise: a case study helps me develop the model framework.

• Feb. 2012: questions about the Progressive Stack.

• March 5, 2012: March is my month of democracy math: planning for the month.

• March 8, 2012: Issues vs. Proposals.

• March 11, 2012: I present this project to Occupy Oakland's Research Working Group, and take notes on the discussion.

• March 12, 2012: Forks in the road: Notes on multiple directions forward.

• March 14, 2012: forks in the road; constraints; complications: An update on my progress, outlining a roughly two-pronged approach.

• March 15, 2012: I present this project at the monthly 5 Minutes of Fame event at Noisebridge, the hackerspace in San Francisco, where this month the 5 Minutes of Fame is also the kickoff for a weekend unconference on horizontal organizing.

• March 20, 2012: slow process, nogoods, disruption: Further notes on changing preferences and structural inequalities, and reflections on the local Occupy organizations.

• March 31, 2012: pausing; negotiation; persuasion; silence: A multistage negotiation process, what happens to people who can't speak freely, and more on changing preferences.

• August 22, 2012: I present my progress on this project at Google.
  • Slides are here on the wiki, and the video of the talk is online here.

Map

• Consensus Experiments documents the details of the research and summarizes some results.
• Consensus Code is the home page for the simulations' source code.
• Results are mostly on Consensus Project Results.
• Consensus Dynamics Notes is formative notes on the project, and ideas for things to do in the future.

To keep track of updates in this project but not in the rest of the wiki, you can use the recent changes in category "Consensus" (rss). There is also the "consensus" category at my lab notebook website, and various other ways to follow things including twitter, facebook, etc.