Background: The shortage of skilled workers choosing STEM (Science, Technology, Engineering, and Math) careers in the USA and worldwide has fueled a movement towards STEAM, in which the "A" addresses the arts and humanities. STEAM education has been proposed as a way to offer relevant problems to solve while drawing on creative and collaborative skills to increase interest and engagement in these fields. Despite the interest in increasing STEAM globally, research on the efficacy of instructional approaches, including ways to assess collaborative problem solving (CPS), is lacking. Results: This paper reports the development of a rubric, named Co-Measure, for researchers and educators to use to assess student collaboration, at the individual level, when students are working in K-12 STEAM activities. Our project team provides the rationale, process, validation, initial iterations to the rubric, and next steps to inform STEM researchers and move STEAM instruction and learning forward. A final rubric is provided and made freely available to researchers and educators. Conclusions: As STEAM education gains popularity in K-12 schools, assessment of student collaboration is needed to identify the dimensions of the skill in order to provide appropriate problem solving opportunities within instruction. The assessment is also needed to adjust instruction when students are not offered opportunities to work collaboratively during STEAM activities. Utilizing existing generalized frameworks of CPS provided the initial guide to direct research specific to CPS in STEAM activities. Using an iterative process to identify and evaluate attributes of student behavior associated with CPS in STEAM classrooms by a project team comprised of learning scientists, educational researchers and psychometricians allowed for rigorous research while drawing on appropriate expertise. Co-Measure has the potential to be modified and broadly extended to assess CPS in elementary and post-secondary classrooms using STEAM instructional practices.