Data SGP is an analysis tool for longitudinal student assessment data that generates statistical growth plots showing students’ academic progress relative to their peers. Data SGP is a useful tool for educators and policymakers alike as it provides information about students’ progress in terms of percentile rankings that are easily understood by the general public. In addition, the use of SGPs in conjunction with state achievement goals and targets allows schools to identify and focus on those students who are most likely to not be on track for meeting proficiency standards within their required timeframes.
SGP is not perfect, however, and requires careful data preparation in order to be utilized effectively. In fact, most problems encountered when using SGP analysis revert back to issues with data preparation. As a result, OSPI recommends that districts begin by learning about SGP through the online data tools available on the BAA Secure Site (M-STEP, SAT).
Once schools have familiarized themselves with the SGP data and tools, they should consider how to use this data to support their accountability systems and educator evaluation programs. SGPs provide a valuable tool for communicating the complexity of state achievement standards and illustrating how schools can accelerate programs in order to reach these standards.
In addition, SGPs are an important component of the college and career readiness performance index and serve as one of several indicators in the teacher and leader key metric systems. To learn more about these uses of SGP please see the Accountability page.
SGP data can be downloaded via the BAA Secure Site under “Reports” and is used in the online data tools created by Macomb and Clare-Gladwin school districts. Once the data is imported into these tools, teachers can then select a specific student and see the student’s growth in both M-STEP and SAT achievement.
While SGPs are based on statistical models and are not a direct measure of teacher effectiveness, many policymakers prefer them over other popular statistical measures such as value-added modeling, because they are more intuitive for teachers and the community. SGPs rank students with similar baseline academic performance in terms of their percentile ranking and are immediately understandable to everyone.
It is important to note that the distribution of SGPs across a school or district will not be a bell curve, as shown in the diagram to the right. Rather, the distribution is expected to be closer to a decile or grouping of 10 percentiles where each percentile contains roughly the same number of students.
In order to run SGP analyses, the data set sgpData must be supplied with 7 required variables: VALID_CASE, CONTENT_AREA, YEAR, STUDENT_ID, CLASS, SCALE_SCORE and GRADE_LEVEL. Additionally, sgptData_LONG or sgptData_WIDE must be supplied with an additional set of variables: DISTRICT_ID, INSTRUCTOR_NUMBER and TEACHER_NUMBER which are required for higher level SGP functions like abcSGP and prepareSGP. In addition, the sgptData_LONG and sgptData_WIDE sets include demographic/student categorization variables which are required for creating teacher level aggregates by the summarizeSGP function.