Preschool Problem Solvers: Developing Assessment Tasks to Measure Young Children’s Learning of Computational Thinking Skills and Practices

Research that explores which computational thinking (CT) skills resonate with the abilities and interests of preschoolers is limited. Even more limited is the availability of assessments that can be used to measure young children’s CT learning. This study describes the process employed to generate a series of developmentally appropriate assessment tasks designed to draw inferences of young children’s learning of CT components and engagement in CT practices. Assessment tasks were designed and pilot-tested with a sample of 57 preschool children in Virginia and California. Item responses were coded and analyzed using classical test theory to examine item difficulty and discrimination. The reliability of the final set of items was calculated using Cronbach’s alpha. The overall difficulty of the items was relatively high, and discrimination values were generally adequate. The total summated score was also found to be reliable. Our assessment and design process reveals challenges and shares ideas for item formats that allow gathering evidence of young children’s CT learning.