Title: Pathways to proposing causes for unexpected experimental results

Abstract: Models of physical systems are used to explain and predict experimental results and observations. When students encounter discrepancies between the actual and expected behavior of a system, they revise their models to include the newly acquired observations, or change their apparatus to better represent their models. The Modeling Framework for Experimental Physics describes the process of matching measurements and observations to models by making revisions to resolve discrepancies. As part of a larger effort to create assessments of students' modeling abilities in the context of upper-division electronics courses, we used the Modeling Framework to develop and code think-aloud problem-solving activities centered on troubleshooting an inverting amplifier circuit. We observed that some participants iteratively and continuously made measurements and revisions if they could not immediately propose a cause for an observed discrepancy. This pathway has not been previously discussed in the Modeling Framework. In this paper, we discuss two episodes where students undergo this process to converge on a proposed cause post hoc. We conclude by discussing implications for a modeling assessment based on the observed modeling behavior.