Although TAs are responsible for a significant portion of students' instruction at many universities, science TAs and their teaching have not been the focus of a significant amount of study. This dissertation begins to fill this gap by examining physics graduate students who teach discussion sections for introductory courses using tutorials, which are guided worksheets completed by groups of students.

This analysis contends that considering the broader influences on TAs can account for TA behavior. Observations from two institutions (University of Colorado, Boulder and University of Maryland, College Park) show that TAs have different valuations (or buy-in) of the tutorials they teach, which have specific, identifiable consequences in the classroom. These differences can be explained by differences in the TAs' teaching environments. Next, I examine cases of a behavior shared by three TAs, in which they focus on relatively superficial indicators of knowledge. Because the beliefs that underlie their teaching decisions vary, I argue that understanding and addressing the TAs individual beliefs will lead to more effective professional development. Lastly, this analysis advocates a new perspective on TA professional development: one in which TAs' ideas about teaching are taken to be interesting, plausible, and potentially productive.