While consistently related to positive teacher behaviors and student outcomes, teacher efficacy as a construct is currently the subject of debate. The present study examined a new model of teacher efficacy (Tschannen-Moran, Woolfolk Hoy, & Hoy, 1998) that proposes to more clearly define important variables and integrate two theoretical traditions in the study of teacher efficacy. A new instrument was developed to assess a means-end task analysis and context specific efficacy, both important parts of the model. Task analysis and both global and context specific efficacy were measured in 109 emergency certification teachers since task analysis may be more explicit for novice teachers. Factor analyses of the global and context specific efficacy measures suggested subtle but important distinctions in constructs related to efficacy. Canonical correlation analysis indicated the importance of personal teaching competence to instructional efficacy judgments and external locus of control to classroom management efficacy. Although task analysis was not found to be a significant predictor of either global or context specific efficacy, exploratory results suggested potential value of task analysis in future assessments of teacher efficacy. General support was found for the Tschannen-Moral et al. model. (Contains 32 references.) (Author/SM)