Game-Based Programming Challenges: Stealth Assessment of Student Competencies

• In Event: Poster Session 10
  In Poster Session: Learning in Immersive Environments: Science, Statistics, Assessments, and Learner Characteristics

Abstract

Game-based learning environments show promise for engaging middle grade students, promoting understandings of Computer Science concepts, and for implementing stealth assessment of student competencies. One challenge in designing these assessments is determining which trace data are informative and scalable. Evidence-centered assessment design was used to guide a strategy for examining trajectories of game performance. Latent Class Growth Analysis identified three distinct trajectory classes of programming performance (measured as time spent programming). Distal outcomes of post-test scores and self-reported engagement were significant predictors of class membership and add to the explication of how three clusters of student pairs progressed through a series of challenge-based programming tasks involving the concept of iteration. This approach shows considerable promise for informing stealth-based assessments.

Authors

• Megan Frankosky, North Carolina State University

• James Creager, North Carolina State University

• Eric N. Wiebe, North Carolina State University

• Philip Buffum, North Carolina State University

• Wookhee Min, North Carolina State University

• Kristy Boyer, North Carolina State University

• Bradford Mott, North Carolina State University

• James Lester, North Carolina State University