Over the course of instruction, not only does most introductory physics students' content knowledge evolve but their attitudes and approaches to problem solving are also likely to evolve. While changes in epistemology and beliefs about physics have been examined in the literature, how students' attitudes and approaches to problem solving change from the beginning to the end of instruction in introductory physics and how method of instruction or gender of the student impact them remain largely unexplored. To examine the potential changes in attitudes and approaches to problem solving over a semester, we administered a previously validated Attitudes and Approaches to Problem Solving (AAPS) survey at both pre-and-postinstruction in eight large enrollment calculus-based introductory physics classes at a large research university in the U.S. At both points in time (beginning and end of a semester), each class was also given surveys measuring students' conceptual understanding: the Force Concept Inventory (FCI) or the Conceptual Survey of Electricity and Magnetism (CSEM). In addition, final exam scores, gender of students, and descriptions of the instructional methods used for each class were collected. We examined students' performance on the AAPS survey, FCI or CSEM, and final exams, and compared the results for different instructional methods, and gender of students. We also examined whether or not there were correlations between the expertlike response on the AAPS survey and the performance on FCI or CSEM or final exams. We found that all classes exhibited a decline in score on the AAPS survey suggesting worse attitudes related to problem solving after instruction. Classes which involved significant use of evidence-based active engagement methods exhibited statistically significantly better scores on the postinstruction AAPS compared to classes taught using lecture-based approaches. Female students exhibited less decline in AAPS scores than males in all classes.