In working towards unraveling the mechanisms of productive collaborative learning, dual eye tracking is a potentially helpful methodology. Dual eye tracking is a method where eye-tracking data from people working on a task are analyzed jointly, for example to extract measures of joint visual attention. We explore how eye gaze relates to effective collaborative learning and how analysis of dual eye-tracking data might enhance analysis of other data streams. In this chapter, we identify three broad areas of analysis where dual eye tracking may enhance understanding of collaborative learning processes: (1) how eye gaze is associated with other communication measures; (2) how eye gaze is associated with features of the task environment; and (3) how eye gaze relates to learning outcomes. We present analyses in each of the three areas through joint visual attention, using a dataset of 28 4th and 5th grade dyads working on an Intelligent Tutoring System for fractions. By combining eye tracking, dialogue transcripts, tutor logs, and pre/post data, we show the potential of using dual eye tracking to better understand the collaborative learning process. [This is pre-publication version of a chapter published in: "Innovative Assessment of Collaboration," edited by P. Kyllonen, A. von Davier, and M. Zhu, New York: Springer, 2017, pp. 157-172.]