The data sgp package is a set of classes, functions and data that are used to conduct operational analyses of student growth percentiles (SGPs) and student growth projections/trajectories using large scale, longitudinal education assessment data. The package utilizes quantile regression to estimate the conditional density associated with a student’s achievement history and generates SGPs and projections based on these coefficient matrices.

Educators are awarded SGP scores based on their performance in 2023-2024 and the median of their previous two years of testing. The Wisconsin Department of Education (MDE) strongly recommends that educators not use their SGP score for high-stakes educator evaluation purposes until the 2020-21 school year at the earliest.

SGP scores provide teachers with a detailed picture of their students’ achievement that can be used for a variety of purposes including planning, instruction, and educator evaluation. Teachers should review their own SGP scores with their data teams and use these insights to identify areas where they need to improve. SGPs can also be shared with parents as a means of enriching the understanding of their child’s academic progress.

A student’s SGP is a number between 1 and 99 and indicates the extent to which their assessed score has increased or decreased in relation to their academic peers. An SGP of 85 means that a student scored higher than or equal to 85 percent of their academic peers.

SGPs are calculated for students in grades 4-8 who have taken at least two grade level assessments of the same content area. For example, a student who took the Badger Exam in grade 7 and the Forward Exam in grade 8 will receive an SGP for both ELA and Math.

The sgpData table contains the student-level test results from all the tests a student has taken in grade 4, 5, 6, 7, and 8, including the SGP score from each of these assessments. The table contains the student’s unique ID and the assessment scores for each of these years. In addition, the sgpData table contains the insturctor lookup file, sgpData_INSTRUCTOR_NUMBER, which identifies the teacher for each student’s test record.

A key to SGP calculations is the fact that students are compared to their peers, not to the mean or average of the group. As a result, even students who typically have high scores on state assessments may have SGPs below 50. The lower SGP in these cases is due to either slight misfit in the norms or the assignment of Highest Obtainable Scale Score (HOSS) students to an SGP of 99.

SGPs can be computed for students at the classroom, school, district, and state levels. Each of these analyses requires the use of an exemplary dataset. The sgpData_LONG, sgpData_INSTRUCTOR_NUMBER, and SGPstateData tables are provided in the Demonstration_SGP vignette along with a short script to run the required SGP analysis. Running these analyses will require a computer that has access to the open source statistical software program R. This program is available for Windows, OSX and Linux and there are many resources on CRAN to help new users get started with it.