The focus of this research was to investigate how students learn to do fieldwork through observation. This study addressed the following questions: (1) Can mobile eye-tracking devices provide a robust source of data to investigate the observations and workflow of novice students while participating in a field exercise? If so, what are the strengths and limitations of mobile eye tracking? (2) If these devices offer a unique source of data to investigate student work, what findings might be helpful for improving field instruction? To address these questions, we used mobile eye-tracking devices in a pilot study to collect video from students completing mapping exercises during a geology field course. Data were collected from six students participating in two different parts of an exercise where they were asked to create geological maps of an area based on their field observations. From this study, we learned that conducting eye-tracking research in field conditions is technically demanding and operationally difficult. We found that most of our analysis was based on reviewing the scene video and did not require the eye-tracking information. In reviewing the scene videos, substantive features of students' observational practices were exposed. We found that students struggle with foundational mapping practices, miss opportunities to collect key data, and are often distracted or disengaged during direct instruction. We observed instances of swarm behavior where students tend to group around outcrops even when nominally working independently. We also noted key differences in student behavior working individually compared with group mapping. We believe these findings provide data for geoscience educators to consider when thinking about ways in which to develop the observational skills of their students and to design appropriate field course instruction.