Data sgp leverages longitudinal student assessment data to produce statistical growth plots (SGP) that measure students relative progress compared with their academic peers. SGPs are calculated for each student based on his or her history of test scores and are reported for both teachers and students. SGPs are useful as they help educators and families understand what their children have been learning throughout the year and can be interpreted much like standard achievement scores.
In Michigan, SGPs are used in teacher evaluation systems along with a variety of other measures of student performance. The SGP data released to districts on the BAA Secure Site in January of each year allows for educators to familiarize themselves with this new assessment information. Michigan law does not require the use of SGPs for educator evaluation in 2015/16 as the state seeks to stabilize the data prior to high stakes use in 2018/19.
SGPs are a useful tool for evaluating teacher effectiveness because they provide teachers with a clear picture of their students’ performance and how their teaching is contributing to or detracting from that performance. SGPs are a useful alternative to traditional growth-adjusted value-added models which are used to evaluate teachers in a more granular manner.
A key difference between the two measures is that SGPs are calculated based on the historical growth trajectories of Star examinees while value-added models are based on the results of the most recent assessments for each student. Using these different calculations leads to differences in the results and interpretation of both measures.
SGPs utilize historical standardized test score data to determine what level of progress a student should be making by comparing his or her most recent assessment to the average performance of students who have taken the same exam in previous years. These historical trajectories are then used to project future performance for the current year. This information is then compared with the performance of other students to determine how much progress the student needs to make in order to reach their target proficiency level.
While SGPs are very informative, the complexity of the calculations involved makes them prone to large estimation errors that can result in inaccurate conclusions about student performance. Therefore, it is important for users to spend a significant amount of time on data preparation prior to running any SGP analyses. In fact, almost all errors that result from running SGP analyses can be attributed to incorrect or incomplete data preparation.
The data sgp package provides users with an easy-to-use script for analyzing SGPs in the open source software environment R. R is available for Windows, OSX and Linux and there are numerous online resources to assist newcomers in getting started with this statistical data analysis tool. The data sgp package also provides exemplar WIDE and LONG formatted data sets (sgpData and sgpData_LONG) to support data preparation for SGP analyses. The sgpData_LONG data set includes a variable sgpData_INSTRUCTOR_NUMBER which provides an anonymized lookup table that links instructors to their students for each content area.