This article reevaluates the latent structure of the Teacher Efficacy Scale using confirmatory factor analyses (CFAs) on a sample of preservice teachers from a public university in the U.S. Southwest. The fit of a proposed two-factor CFA model with an error correlation structure consistent with internal/ external locus of control is compared to the historical CFA model. A majority of the historical problems with items not properly loading and/or with items cross-loading are avoided in the proposed model. Moreover, it exhibits superior fit and is more consistent with the theoretical foundations of the Teacher Efficacy Scale than past CFA models. Additionally, the effect of the internal/external locus-of-control dimension of the scale is confounded with the methodological effect of positive and negative wording. Future remedies for these confounding effects are discussed.