Abstract Spatial skills, which represent the ability to visualize and imagine manipulating objects in one’s mind, are necessary for success in the science, technology, engineering, and mathematics (STEM) fields and are particularly relied upon by geoscientists. Although scholars recognize the importance of these skills, explicit training is inconsistently offered throughout courses. Furthermore, the relationship between spatial training and students’ perspectives on STEM fields is underexplored. To address this, we developed a case study that included over 700 students enrolled in introductory geology classes over three semesters. These students were randomly divided into control and experimental groups; the experimental group completed 10 spatial training assignments, and the control group completed the course as usual. We relied on situated expectancy-value theory to interpret changes in students’ perceptions of both the course and science overall, and asked the following research questions: (1) Do students who complete the spatial training assignments (i.e., the experimental group) have a statistically significant improvement in their final course grade, self-efficacy, and/or value when compared to the control group? (2) If so, what are the effect sizes of these changes? (3) Is there a minimum number of trainings that need to be completed to achieve this effect? (4) Is there a maximum number of spatial training assignments where we stop seeing improvement (ceiling effect)? We surveyed all students enrolled using 38-question pre- and post-assessments of their self-efficacy, defined as a belief in their ability to succeed, and value. We found that between the control and experimental groups, there were significant differences in students’ pre- to post-changes in perception of science self-efficacy, class self-efficacy, and class value. We found non-significant between-group differences in final grade and science value. We interpret this to mean that using weekly spatial training assignments could increase students’ perceived self-efficacy in their introductory geology course as well as in science more broadly, potentially having ripple effects that support students’ long-term engagement with the sciences. Findings suggest that practitioners should include explicit spatial training in their courses to improve students’ perceptions of the course and science overall. Building on this work may include describing to students the purpose of spatial training (which was deliberately avoided in this study) and outlining the research that supports the relationship between spatial skills and success in STEM fields. Future directions may also include longitudinal tracking of spatial and related skill development throughout students’ college careers.