Description: Students often pursue one of two achievement goals: mastery goals, which focus on learning and skill development, and performance goals, which focus on outperforming peers and earning high grades. Both can be framed in an approach manner (to strive for success) or an avoidant manner (to avoid failure), but the approach forms are of greater theoretical interest here. Because these two goals represent different ways of pursuing competence, theorists posit that they should produce distinct educational experiences. In particular, it has traditionally been assumed that mastery goals are more adaptive and beneficial, producing educational outcomes similar to or better than performance goals, but never worse (Dweck & Leggett, 1988; Nicholls, 1984). However, many studies, particularly in educational settings, have instead shown distinct benefits of both goals: across various ages and types of classes, mastery goals typically predict course interest but not achievement (i.e., grades), whereas performance-approach goals typically predict good grades but not interest (see Harackiewicz, Barron, Pintrich, Elliot, & Thrash, 2002). Theorists are currently at a loss to explain the surprising disconnect between mastery goals and grades, especially in light of the performance-approach goal's favorable link to grades and the fact that mastery goals often predict deep learning strategies (e.g., elaboration of course material and questioning its veracity), which one would expect to translate into better class performance. We propose a novel explanation for this disconnect between mastery goals and grades: while studying, students pursuing mastery goals might sometimes undermine their class performance by pursuing their own learning agenda instead of the teacher's agenda. They might, for instance, delve deeply into a personally interesting subset of course material, or even explore interesting tangents to the material, to the neglect of duller material that the teacher deems important. Consequently, they may find themselves under-prepared for exam questions testing those duller topics; hence why this goal does not reliably predict grades. Performance-oriented students, due to their preoccupation with earning high grades, should be less apt to pursue their own learning agenda and, therefore, also be less likely to sabotage their own performance in this manner. This study provides an initial test of this "tangential studying" premise. Methodology: 263 General Psychology students completed two questionnaires. The first, administered the week before the first exam, measured students' achievement goals for the class. The second, administered before the last exam, assessed their interest in the course material, their usage of surface versus deep learning strategies, and their engagement in Tangential Studying (6 items; e.g., "While studying, I often spend more time reading things I find interesting than what is required"). Conclusions: Regression analyses tested and supported three hypotheses. First, replicating prior research, mastery goals predicted course interest (ß = .46, p < .001) but not grades, whereas performance-approach goals predicted high grades (ß = .19, p < .05) but not interest. Second, although mastery goals predicted use of deep learning strategies, these did not in turn elevate grades. This too replicates many prior studies. Third, and most important, mastery goals also predicted engagement in tangential studying (ß = .21, p < .01), which in turn did predict lower course grades (ß = -.16, p < .05). In sum, these findings suggest that mastery-oriented students allocate study time disproportionately to personally interesting course material, often neglecting duller material, and therefore risk under-achievement on the exam. By contrast, performance-oriented students were unlikely to engage in tangential studying (ß = -.16, p = .05), perhaps due to their high adherence to the teacher's learning agenda. Nor, then, did they inadvertently sabotage their own exam performance. Implications for goal theorizing are discussed.