Background: Although teacher-student relationships lie at the heart of students' schooling experience, fundamental questions regarding these relationships remain unanswered. Aims: This study investigates three related questions about these relationships: To what extent do they change from the beginning to the end of a school year? Are any emergent changes associated with shifts in students' academic or motivational outcomes? Are certain "upstream" factors associated with improvements or declines in teacher-student relationships? Sample: We investigate these questions with a sample of middle school students (N = 119) and their teachers (N = 30). Methods: Through a novel approach which accounts for both perspectives within teacher-student relationships, we assess these relationships at the beginning and end of the school year. Using multi-level models, we examine how changes in these relationships are associated with changes in students' grades, homework completion rates, self-efficacy, and effort. In addition, we examine associations with two potential precursors to teacher-student relationships: students' accuracy in taking their teachers' perspective and their perceptions of similarity to their teachers. Results: We find that substantial changes occur in these relationships from the beginning to the end of the year; these changes are associated with shifts in important student outcomes; and changes in students' social perspective taking accuracy and perceived similarity to their teachers correspond with changes in teacher-student relationships. Conclusions: Given the malleability of teacher-student relationships and their importance for key achievement and motivational outcomes, we advocate for researchers to conduct field experiments to inform how to improve these critical relationships. (Contains 3 footnotes and 4 tables.)