The SGP package includes classes, functions and data used to calculate student growth percentiles and percentile growth projections/trajectories using large scale, longitudinal education assessment data. It uses quantile regression to estimate the conditional density associated with each student’s achievement history and derived coefficient matrices to produce projections/trajectories that show the percentile growth needed for a student to reach future achievement targets.

While student growth percentiles are useful, they should not be the sole factor when making educational decisions. Students should be assessed on various metrics, including standardized tests and classroom performance evaluation. Additionally, educators must consider individual student SGPs when making decisions about how to support or challenge students as well as using SGP results in curriculum planning processes for future instruction.

In addition to the SGP packages, a number of other R libraries and functions may be required for some analyses. For example, the lmgrd and tbrrd libraries are needed for some statistical operations. In addition, the sgpData library is required for creating state level SGP data sets. The sgpData library contains the code and functions for managing, loading, and analyzing data set definitions, SGP calculations, and results.

Educators should also be aware that the SGP models have limitations, including the potential for spurious correlations due to differences in school/teacher characteristics and baseline cohort design. In addition, the models can be misleading for students that already possess superior academic abilities. For these reasons, it is recommended that educators review their SGP reports with an experienced specialist before distributing them to families and teachers.

SGP analyses require the use of a computer that has the free open source software, R, installed on it. R is available for Windows, OSX, and Linux and can be downloaded from the CRAN repository. While there are many resources available to get started with R, it is recommended that users familiarize themselves with the software before diving into running SGP analyses.

The SGP packages can be run in a number of ways, with varying degrees of complexity. The lower level functions studentGrowthPercentiles and studentGrowthProjections can be run separately or combined into a single, comprehensive function called prepareSGP. prepareSGP takes exemplar LONG format data, sgpData_LONG and state specific SGP meta-data, sgpData_INSTRUCTOR_NUMBER. The function then produces an SGP object, Demonstration_SGP, that displays the results of the analyses.

For operational analyses, it is recommend that the data be saved in the LONG format (sgpData_LONG). This makes it easy to add additional years of data without rerunning all of the SGP functions and generating new output. In addition, the higher level functions often assume that the data exists in the LONG format and do not need to be rerun. This simplifies the process of updating analyses with an additional year of data and makes the SGP package more robust and user friendly. However, users can still choose to save their analyses in the WIDE format if desired. This allows them to easily compare the results of different analyses conducted with the same data.