Teacher performance scores inform education leaders' management of teacher human resources. However, prior research has implied that different interpretations of performance criteria between teachers and their evaluators suppress teacher development. Although research has examined teacher perceptions of performance scores and compared teacher self-scores with researcher-generated scores, no work has estimated the statistical differences between self-scores and field-based evaluator scores. I applied structural equation modeling to a unique dataset of over 4,000 teacher self-scores and 4,000 evaluator scores of teacher performance to examine the extent to which teachers and evaluators applied a standards-based teaching performance rubric similarly. Consistent with prior work, some average self-scores exceeded average evaluator scores; however, evaluator-assigned scores exceeded self-scores on one indicator. The evidence also suggested that teachers and evaluators interpreted the meaning of two performance indicators differently. The findings imply that there are opportunities to strengthen evaluators' and teachers' shared understanding of teacher performance criteria. Teachers and evaluators might develop shared understandings by viewing videorecorded lessons and collaboratively interpreting teaching episodes using the performance criteria. Education leaders might also emphasize that preobservation conferences are an opportunity for evaluators and teachers to discuss performance criterion interpretations. As some selfscores exceeded evaluator scores, education leaders could introduce a consider-the-opposite scoring strategy, where evaluators consider two reasons that the score they want to give could be inaccurate, prompting stronger scoring justification, potentially leading to more accurate scoring.