Since the introduction of Tanzania’s secondary education expansion policy in 2004 student performance has declined significantly. This trend led me to hypothesise that there may have been a decline in teacher efficacy. The present study investigated the influence of the policy and its implementation on teachers’ self-efficacy beliefs, because these beliefs impact on the quality of teaching. The study used a mixed-method approach, with a focus on case studies. Interviews, questionnaires and documentary reviews were used for gathering data. Ninety-nine secondary school teachers from four schools in the Iringa region of Tanzania participated in this research. Social cognitive theory was employed as a conceptual framework for the project. While inductive procedures were used for analysing the interview data, the questionnaire data were analysed via the Statistical Package for Social Sciences Software (SPSS) before being tabulated and graphed. Findings indicated that strategies intended to ensure the successful enactment of the expansion policy were not fully implemented nor targets met. In a number of ways this adversely affected the ability of teachers to go about their work and, consequently, educational outcomes for students. Overall, evidence from this study suggests that the manner in which the expansion policy was enacted and implemented contributed to the erosion of teacher professional identity and social standing, particularly in relation to self-efficacy. This produced further vicious cycles of effects and consequences. Factors associated with school staffing, school infrastructures, resources for teaching, relationships with the community and poor academic standards of enrolled students emerged as central contributors to this negative situation. An analysis of quantitative data strongly suggested that the expansion policy was implemented poorly, which was detrimental to teacher efficacy and, ironically, undermined the policy itself. This research adds to a growing body of literature on the sources of teachers’ self-efficacy. Firstly, the study identifies evidence which indicates that social persuasion is one of the most important contributors to social self-efficacy. It also identifies factors which reinforce a view of mastery experiences as the most influential source of self-efficacy. Secondly, the study points out the causes of teachers’ negative emotional states, including factors outside of the school environment. Thirdly, the study found that certain factors which were emphasized by Bandura as having impact on only one specific source of self-efficacy actually have an impact on more than one source of self-efficacy. Analysis of the research findings generated three major policy recommendations for helping teachers develop self-efficacy: promoting professional learning; empowering parents and education authority figures; and improving teaching and learning environments. The study also indicates areas for further research in factors related to teacher efficacy in Tanzania. These include: teaching practices in new schools created under the policy; the extent to which the curricula of university teacher education courses influence self-efficacy in professional practice; and how the policy impacts upon the self-efficacy of students as learners.