Galileo: Citizen-led Experimentation Using a Social Computing System
Vineet Pandey, Tushar Koul, Yang Chen, Daniel McDonald, Mad Price Ball, Bastian Greshake Tzovaras, Rob Knight, Scott Klemmer
Abstract
People have scientific questions and folk theories; yet most lack the expertise to investigate them. How might people transform their questions into experiments that inform both science and their lives? This paper demonstrates how online volunteers can collaboratively design and run experiments using a novel social computing system. The Galileo system provides procedural support using three techniques: 1) experimental design workflow that provides just-in-time training; 2) review workflow with scaffolded questions; and 3) automated routines for data collection. We present two empirical investigations: a study and a field deployment with online volunteers across 16 and 8 countries respectively. People generated structurally-sound experiments on personally meaningful topics; three communities ran a week-long experiment each. We identify two key challenges for citizen-led experimentation—supporting different expertise levels and providing recruitment guidance—and provide specific suggestions from the social computing literature. Our results highlight the promise and challenges of citizen-led knowledge work like experimentation.