Policymakers are increasingly adopting "pay-for-performance" policies in which teachers are compensated based on their performance as measured by classroom evaluations and/or student achievement test results. Prior research has produced largely inconclusive findings concerning support among teachers for these policies and their effects on teachers. This paper analyzes teacher survey data drawn from a larger study of four high-poverty, urban school districts as they implemented a pay-for-performance initiative. The paper reports on attitudes about compensation reform in general, support for and perceived benefits and drawbacks of the policy in question, as well as on perceived effects the policy on teachers' working conditions, specifically in the areas of collaboration. Findings indicate high levels of support for pay-for-performance in general but somewhat less support for specific components of the program, such as using student performance to determine teacher pay. Responses were largely mixed regarding the effects of the policy on working conditions such as collaboration. A quasi-experimental design was used to allow for the comparison of responses on a select set of survey items, and teachers' participation in the program generally had no effect on their attitudes about compensation when compared to teachers in similar schools who had not participated in the program. Differences in survey responses based on teacher-level characteristics are also addressed, but these did not explain variations in teachers' responses. Despite teachers' generally favorable perceptions of the pay-for-performance program, their knowledge of the program's details was limited throughout the period of implementation, thus raising serious questions about the program's ability to achieve its goal of bringing about change in teacher behavior. An appendix provides additional information about the procedures used to conduct the factor analysis. (Contains 7 tables and 9 footnotes.)