Teacher evaluation systems treat instructional quality as generic. Principals often observe elementary teachers in one subject and generalize assessments to all subjects. However, there is little empirical work to justify these decisions. This study provides needed evidence about whether elementary teachers engage in comparable instruction across the school day. We draw on data from the Measures of Effective Teaching (MET) project, including student survey and classroom observational data from more than 500 elementary teachers. Findings indicate that there is moderate within-teacher, cross-subject consistency on the Tripod and Classroom Observation Scoring System (CLASS). Cross-subject correlations are higher on the Tripod scales (r values from 0.55 to 0.73) than the CLASS dimensions and domains (r values from 0.25 to 0.55). These findings suggest that teaching quality is not a uniform construct across subjects, even though current teaching evaluation systems largely treat it as such. Implications for elementary teacher preparation, professional development, and evaluation are discussed.