Abstract: Collaboration and group work are highly-valued classroom practices, and they have featured prominently in current
initiatives concerning the measurement of “21st century skills.�� However, fundamental questions remain about how
to design and model assessments of collaboration. In particular, while self-reports and situational judgment tasks
can be useful research tools, they are not ideal for making consequential decisions about students in educational
settings. One way to address this issue is through the use of performance-based assessments, which can provide
observable evidence about student competencies. To this end, each paper in this symposium presents novel
research on the performance-based assessment of collaboration. The first paper emphasizes design principles
based on recent theoretical work on how students learn and solve problems in collaborative settings. The second
paper analyses data from three novel task designs using a multi-group IRT-based approach. The third paper studies
response time distributions in a set of simulation- and collaboration-based assessments that involve students
interacting to solve the same problem. The fourth paper addresses how to simulate data from complex interacting
systems, in order to analyze the psychometric properties of models used to study collaboration.

Keywords: collaboration assessments, data simulation, interacting systems, performance-based assessments