Comparing Cognitive and Situated Assessments of Learning in Middle School Computer Science

Abstract

As K-12 Computer Science becomes a mainstream subject, there has been recognition of a need for dialogue between various theoretical framings of learning. However, even in research showing the importance of sociocultural factors, quantitative assessment of Computer Science learning has been predominantly cognitive. This study presents an example of how quantitative assessments of learning, based on cognitive and sociocultural framings of learning, can be put into dialogue by developing measures of learning on internal individual terms and understood as participation in a community of practice. We develop two participation-based constructs assessed using methods from learning analytics, and show that each is significantly associated with better performance on a cognitively-based summative assessment of computer science content. These associations are mediated by the content of students' programs. Beyond serving as contextual factors for cognitive assessments, we propose treating these constructs as primary evidence of learning.