The Texas Teacher Evaluation and Support System Rubric

A new study from Regional Educational Laboratory Southwest examines the statistical properties of the Texas Teacher Evaluation and Support System (T-TESS) rubric from the pilot year. In 2014-15 the Texas Education Agency piloted the T-TESS in 57 school districts. During the pilot year teacher overall ratings were based solely on rubric ratings on 16 dimensions across four domains. The study examined the statistical properties of the T-TESS rubric to explore the extent to which it differentiates teachers on teaching quality and to investigate its internal consistency and efficiency. It also explored whether certain types of schools have teachers with higher or lower ratings. The dataset included over 8000 teachers across 251 schools and 51 districts that participated in the pilot. Descriptive statistics were reviewed to assess the extent to which the T-TESS rubric ratings differentiate teacher effectiveness. Correlational analysis was performed to determine the internal consistency of the rubric. Uniqueness values, which resulted from a factor analysis of T-TESS’s 16 dimensions, were examined to determine whether each dimension makes some unique contribution. Lastly, regression analysis was conducted to explore the relationships between T-TESS performance ratings and school characteristics.  Results indicate that in the 2014/15 pilot of T-TESS, 1.6 percent of teachers were rated as improvement needed, 24.9 percent as developing, 68.3 percent as proficient, 3.7 percent as accomplished, and 1.5 percent as distinguished based on the T-TESS rubric for teacher effectiveness. The T-TESS rubric is internally consistent at both the domain and dimension levels. All dimension-to-dimension within domains and domain-to-domain correlations are positive, suggesting that none of the domains or dimensions stand out as unrelated or contradictory to the rest of the system.  Findings also suggest that the T-TESS rubric is efficient. None of the domains or dimensions are clearly redundant, as supported by findings that no correlation is close to one. In addition, an analysis of uniqueness reveals that each dimension makes some unique contribution.  Although statistically significant relationships are found between observation ratings and school characteristics, the combination of observed student and school characteristics explains, at most, approximately 8 percent of the variation in overall observation ratings for high schools, and explains even less for elementary and middle schools.  One area for future research is the validation of the ratings with other measures of teacher effectiveness, such as student growth. However, the validity of a teacher evaluation system itself may not necessarily translate into improvements in teacher effectiveness or into long-term outcomes, such as teacher retention and greater student achievement. Therefore, further research could explore whether the implementation of such systems do in fact relate to the more distal measures.