Recent Changes - Search:



Bayesian Networks in Educational Assessment

Cognition and Assessment


edit SideBar



The ‘Biomass’ project featured a prototype of an innovative assessment product, and the ECD process that led to the design. The goals of the project were to demonstrate (1) a assessment product designed to serve two new purposes in the transition from high school to college, and (2) the capability needed to produce this kind of assessment product. Biomass was supported from its inception in January 2000 through June 2000 by Educational Testing Service (ETS) and the College Board (CB), through their Assessment Futures joint initiative, and by the Research Division of ETS from July 2000 through its completion in September 2000.

The project instantiates the general ECD framework in terms of the specifics of the purposes, the domain, and the constraints of an intended assessment product: Secondary biology, with a focus on investigation and model-based reasoning, in the domains of microevolution and modes of inheritance. A web-based platform was used for delivery.

The project is documented in two CRESST Research Reports:

Introduction to the Biomass Project: An Illustration of Evidence-Centered Assessment Design and Delivery Capability, CSE Report 609, by Linda S. Steinberg, Robert J. Mislevy, Russell G. Almond, Andrew B. Baird, Cara Cahallan, Louis V. Dibello, Deniz Senturk, Duanli Yan, Howard Chernick, Ann C. H. Kindfield.

This report describes the sequence of activities that led to the requirements for the assessment product, in terms of the objects in the CAF and the processes the assessment delivery system. Examples from the prototype show how these requirements are realized in specific assessment elements. The project led to a working prototype of an assessment product, but one constructed on the lab bench. However, the conceptual framework, the delivery system architecture, and the design process through which the prototype was developed are meant to be scalable.

Specifying and Refining a Measurement Model for a Simulation-Based Assessment, CSE Report 619, by Roy Levy and Robert J. Mislevy.

This report focuses on the use of Bayesian inference networks in Biomass for evidence accumulation, or test-level scoring.

Edit - History - Print - Recent Changes - Search
Page last modified on March 02, 2007, at 07:36 AM