What is Data SGP?
Data sgp is a statistical software package designed to facilitate the analysis of large scale, longitudinal education assessment data. The packages classes, functions and data structures are based on the concept of student growth percentiles (SGP) and percentile growth projections/trajectories which utilizes quantile regression to estimate conditional density for each students achievement history.
The sgp package has higher level functions (wrapper functions) that simplify the source code associated with operational SGP analyses. However, before you can conduct any analysis you will need to have some familiarity with the R software environment. This is an open source, free of charge, programming language that runs on most platforms including Windows, OSX and Linux.
A Student Growth Percentile, or SGP, measures a students growth compared to their academic peers. It allows teachers to see the significant progress a student is making even if they are not yet meeting grade-level standards. This knowledge can inform teachers of opportunities to adjust their instruction and provide support for students in need.
SGPs are reported on both the individual student and class-level reports. An individual SGP score compares a student to all other students with similar prior test scores. An average SGP score compares a students growth to all other students who have taken the state assessment in the same grade and content area. The sgp package provides an mSGP conversion chart to convert a students mSGP from a 1-99 scale to a 1-4 score that can be reported to the teacher in the teacher evaluation system.
SGP analyses use two types of data: (1) a list of a students SGP scores from the state assessments and (2) a set of teacher-student-course relationships. The sgpdata_long data set consists of a year, a content area and an identification for each assessment record along with a student-teacher lookup table sgpData_INSTRUCTOR_NUMBER that associates a teachers ID number with each students SGP score. The NJDOE links district course roster submissions with relevant mSGP scores to determine the teachers of each student for each content area assessment in each school year.
The higher level SGP functions sgpPercentiles and sgpGrowthProjections require the sgpdata_long data set as well as the embedded SGPstateData meta-data. In order to use these functions you will need to have a working copy of the SGPdata_long data set on your computer. This data is available on the Star server and can be obtained by using the report customization function and selecting the sgpdata_long option. The SGP package has wrapper functions abcSGP and updateSGP that will manage the creation and management of these data sets and perform the higher level SGP analyses for you. They are a great way to quickly start evaluating the SGP tool. Just be sure to keep in mind that preparing the data is the bulk of the work for running these SGP analyses. Once the data is prepared the analyses are fairly simple to execute. This is why these are the preferred method for operational SGP analyses.