ITR: Evaluating education -- what are we measuring and how?

Columbia University, New York NY

Investigators

Abstract

The educational system can be viewed as a dynamic environment influenced by many factors. In this project a multi-agent interaction and simulation model will be investigated as a tool for evaluating educational systems as dynamic environments. The long term aim of the tool will be to model the interaction between education policy and human learning by first using typically measured factors such as test scores and economic status and they exploring other factors that are harder to quantify but which may result in a more accurate model. The proposed work involved three phases: (1) constructing an initial model and calibrating it against historical data; (2) examining how other, less typical characteristics can be quantified and expanding the model to include these additional measures; and (3) introducing the model to education policy makers, building a user-interface that they can use and understand; and conducting a pilot study with potential users. The research team consists of computer scientists, dynamical system theorists, multi-agent researchers, school system experts, and AI researchers.