Title: Undergraduate and Graduate Students’ Attitudes and Approaches to Problem Solving Abstract: Student attitudes and approaches to problem solving in physics and astronomy may influence their development of expertise, as well as their engagement and perception of physics and astronomy as an academic endeavor. Introductory physics and astronomy undergraduate classes, which are gateways to a major in physics and astronomy, are foundational experiences in physical science education and development of problem solving skills. Understanding undergraduate students’ attitudes and approaches may shed light on these formative experiences and what may be done to improve such experiences. On the other end of the spectrum, physics graduate students are expected to have developed significant problem solving expertise, and are potential future faculty. In their role of teaching assistants (TAs) and/or in a future capacity as instructors, graduate students may be responsible for making decisions on the types of problems used to shape their introductory students’ experiences. These decisions by TAs may be crucial in the development of introductory student problem-solving expertise. Therefore, graduate students’ attitudes about the instructional merits of different physics problems are worthy of examining in order to inform professional development programs for graduate teaching assistants. Investigating both undergraduate and graduate student perceptions about problem solving, we analyzed data related to gender, course and method of instruction, and type of problems preferred. Our data suggest that female introductory students and introductory students instructed in an evidence-based active engagement manner have more favorable attitudes and approaches to problem solving compared with male students and traditionally-instructed students. Similarly, introductory astrnomy students were found to have more favorable attitudes than introductory physics students. Moreover, it was found that graduate students’ preferences regarding the types of problems they prefer to use with their introductory students does not always reflect the potential instructional benefits afforded by those problems. These findings illuminate pathways toward improving both teaching and learning of problem solving in college physics and astronomy courses.