The Conceptual Survey of Electricity and Magnetism (CSEM) has been used to assess student understanding of introductory concepts of electricity and magnetism because many of the items on the CSEM have strong distractor choices which correspond to students' alternate conceptions. Instruction is unlikely to be effective if instructors do not know the common alternate conceptions of introductory physics students and explicitly take into account common student difficulties in their instructional design. Here, we discuss research involving the CSEM to evaluate one aspect of the pedagogical content knowledge of teaching assistants (TAs): knowledge of introductory students' alternate conceptions in electricity and magnetism as revealed by the CSEM. For each item on the CSEM, the TAs were asked to identify the most common incorrect answer choice selected by introductory physics students if they did not know the correct answer after traditional instruction. Then, we used introductory student CSEM post-test data to assess the extent to which TAs were able to identify the most common alternate conception of introductory students in each question on the CSEM. We find that the TAs were thoughtful when attempting to identify common student difficulties and they enjoyed learning about student difficulties this way. However, they struggled to identify many common difficulties of introductory students that persist after traditional instruction. We discuss specific alternate conceptions that persist after traditional instruction, the extent to which TAs were able to identify them, and results from think-aloud interviews with TAs which provided valuable information regarding why the TAs sometimes selected certain alternate conceptions as the most common but were instead very rare among introductory students.