COMPUTATIONAL SOCIAL SCIENCE

Department of Computational Social Science Seminar

Friday, February 22 - 3:00 p.m.
Center for Social Complexity Suite
Research Hall, 3rd Floor

Using Primary Sources When Creating Models

Wayne Zandbergen
CSS PhD Candidate
Computational Social Science
George Mason University

Abstract: It is often the case that models are constructed without consideration for where, or whether, input data is available. Inputs can be so abstract as to have little relation to empirical sources. The Diamond & Dybvig model of depositor behavior is an example. This presentation will examine developing a model of bank depositor behavior cognizant of potential data availability as well as data limitations. The model utilizes primary source data to populate certain model inputs. Methods of using primary source data to structure and populate the model will be examined, as well as a discussion of limits and challenges of using such sources.