Abstract

Experimental research on problem-solving relies on a wide variety of different tasks. Examples range from artificial low-dimensional problems such as the multi-armed bandit, to high-dimensional real natural science problems such as cooling down atoms. We take departure in recent discussions questioning the utility of canonical tasks for studying problem-solving. The aim is to discuss the potential benefits of a unifying framework and propose a tentative framework for experimental tasks. The talk is an atypical presentation since we want to spur a debate: What are the features of a task you deem essential for studying problem-solving? What is missing in the current experimental paradigm?

Contact

Carsten Bergenholtz, Associate Professor, Department of Management

Kristian Tylén, Associate Professor, Center for Semiotics

Oana Vuculescu, Assictant Professor, Department of Management