Data SGP and Teacher EffectivenessData SGP and Teacher Effectiveness
Data sgp is a new measure of relative student growth that is being used as an alternative to standard test scores in some states. SGP compares a student’s current achievement level against his or her prior achievement level, ranking the student against others with similar previous performance. SGPs are viewed as more fair and relevant to the evaluation of both individual student progress and educator effectiveness than simple rank-ordering of unadjusted achievement levels. SGPs are also viewed as less susceptible to teacher-level bias than simple mean and median rankings, because they take into account both the overall rank of the student and the overall range of prior achievement for students with that same initial level of performance.
SGPs are estimated from longitudinal assessment data, typically using a model that correlates the current achievement level of a student with his or her past achievement level. These estimated SGPs can be correlated with a variety of student background variables, including student gender, race/ethnicity, home language, and school type. As a result, the distributions of true SGPs are expected to vary with student covariates. This variation can have important consequences, such as limiting the ability to use these measures as an indicator of teacher effectiveness.
Our results show that the distributions of true SGPs do indeed vary with student covariates. In particular, schools with a more diverse student population have a wider range of true SGPs for their students than those schools with fewer minority and low-income students. In addition, teachers with a more diverse student body tend to have higher aggregate SGPs than teachers who teach students with fewer minority and low-income students.
These relationships would still exist even if student background variables were perfectly measured through tests that did not suffer from estimation errors, and even if all teachers taught the same students with the same background characteristics. The reason is that these relationships are caused by systematically sorting more or less effective teachers to schools and classrooms that vary systematically with respect to our student background variables.
To help address these issues, we have developed a new utility function to provide SGPs for individual students based on their historical growth trajectories. These new SGPs are reported in two formats: Window Specific SGP, which allows users to compare and report student growth across windows of time; and Current SGP, which is intended to be a quick check-in of a student’s most recent SGP. This functionality is available to all Star examinees who have 5 or more years of assessment data. Please consult the SGP analysis vignette and the WIDE format data set sgpData for more information about how to use these functions and sgpData in general. SGPs are also available for all Star examinees on our web-based dashboard.