The game Shadowspect has been developed as an experimental environment that supports exploration and creativity with assessment principles integrated into the design and development process from the start. Academically, it targets Common Core standards in high school geometry and three-dimensional objects. In addition to the academic standards, the game mechanics are designed to assess persistence, creativity, and spatial reasoning.

Shadowspect is a geometry-based puzzle game that assesses 3D geometry skills, spatial reasoning, creativity, and persistence using movement and manipulation of 3D shapes. It includes two modes: a puzzle mode, where players may try their hand at designer-created levels, and a sandbox mode, where players can build their own levels with the available tools. In puzzle mode, players progress through different puzzles using a limited set of 3D shapes—spheres, cubes, prisms, and so on—to construct figures that match given silhouettes. The objective of each level is to match all three given silhouettes, each showing the same construction pictured from a different angle. Players can move, rotate, and stretch shapes as needed to accomplish their goals.

During gameplay, Shadowspect captures potentially meaningful data and interactions that will be utilized by researchers to build assessment machinery implemented within the assessment system. Because there are multiple ways to solve the more complex puzzles provided in Shadowspect, each individual will have a unique methodology for reaching the same solution, and these rich and varied action sequences are captured as well. Shadowspect accumulates this data to understand what players already know and can do and the diverse variety of ways that they demonstrate their skills. Researchers can use this data to discover patterns of thinking and interaction beyond the targeted constructs being assessed.

Once it is fully developed, we envision Shadowspect being used as an ongoing assessment tool. Geometry teachers might have students play Shadowspect for 5-10 minutes at the end of class, 2-3 times a week. In that chunk of time they could play through a couple of levels, generating enough assessment data for the models to provide regular feedback on their progress, or warnings about the areas they may be struggling with. This information could in turn be fed back into the game, interpreted and shared directly with players, and stream into a data dashboard which teachers could use to inform their student conversations and instructional approaches.