Data SGP offers student growth percentiles and projections/trajectories using large scale longitudinal education assessment data. Student growth percentiles are determined using quantile regression – which compares students with similar score histories to determine relative performance – while projections/trajectories demonstrate what relative improvements need to occur to meet future achievement targets using this same model.
A: Student Growth Performance (SGP) is a statistic which compares students’ current MCAS scores with the scores of academic peers who have taken previous state assessments in the same subject and grade. Based on prior MCAS results, students are placed into one of seven peer groups and their current scores compared with the scores of their peers in that group – this process repeats itself for all tests taken by an individual student.
As SGP is a statistical calculation, it is important to remember that individual student scores will vary slightly as will results for schools, districts and individual teachers. This variation is particularly notable with regard to median SGP scores for individual students which can range anywhere between 50 (+/- 1) to 49 (-/+1). Teachers typically calculate their mSGP score using an average of their Fall 2023-2024 and Fall 2024 scores.
When reviewing a school’s SGP results, it’s essential to remember that these are an average of all teachers teaching the subject for that year. Therefore, teachers in every grade must submit accurate course rosters through NJ SMART so their mSGP score accurately reflects student growth.
However, many will continue to view it as necessary to buy new. So I thought I’d make my case. Though SGP analyses are possible with either WIDE or LONG formatted data, for operational analyses we advise users who wish to run analyses regularly to use our higher level wrapper functions (studentGrowthPercentiles and studentGrowthProjections) with LONG formatted data as these functions assume there exists both an embedded data set (sgpData_LONG) and student instructor lookup table (sgpData_INSTRUCTOR_NUMBER). This approach makes analysis simpler by offering one function call that encompasses all required analysis steps and thus reduces source code requirements for conducting these analyses. Errors often arise during data preparation, so this step will also help reduce errors that might crop up later. Please refer to the SGP state data documentation and sgpData_LONG documentation for more detailed guidance on how to prepare and utilize these functions for operational analyses.