Virtual Performance-Based Assessments

Abstract

Virtual performance-based assessments (VPBAs) are environments for test takers to interact with systems, sometimes including other persons or agents, in order to provide evidence about their knowledge, skills, or other attributes. Examples include tasks based on interactive simulations, games, branching scenarios, and collaboration among students communicating through digital chats. They may be used for summative purposes, as in certification examinations, or for other purposes, as in intelligent tutoring systems and exploratory learning environments. They afford opportunities to obtain direct evidence about capabilities that inherently involve interaction, such as inquiry and collaboration. Our focus here is digital, usually with regard to the environment but always with regard to the form of data. Digital data capture makes it possible to acquire rich details about students’ actions and the evolving situations in which they occur. The challenges they pose to psychometrics lie in designing VPBAs to optimally evoke the targeted capabilities, providing students with affordances that evidence that cognition, capturing the relevant aspects of the performances, identifying meaningful patterns in performances that constitute evidence about the targeted capabilities, and providing an inferential framework for synthesizing the evidence and characterizing its properties. This chapter provides an introduction to VPBAs and psychometric considerations in VPBA design and analysis.