Data SGP (Student Growth Profiles) are aggregated student performance data collected over time that educators and administrators use to assess students, inform classroom practices, support classroom research initiatives and evaluate schools/districts. While standard assessment metrics such as mean or median scores rely on absolute values with fixed scales for scoring purposes, SGP metrics compare individual performance against academically similar peers to provide meaningful percentile scores that help educators better comprehend a student’s progress more meaningfully than standard test scores alone.

Data sgp is a software application designed to analyze large-scale, longitudinal education assessment data and equip its users with tools necessary for calculating student growth percentiles (SGP) and creating student SGP trajectories and projections. SGP analyses use R, an open source statistical programming language free for Windows, OSX and Linux users. While performing SGP analyses requires advanced statistical methods and knowledge of mathematical models underlying it all; numerous resources online exist that provide basic R usage basics as an introduction.

SGP analysis uses quantile regression to generate conditional density matrices for each student’s assessment history and projections/trajectories from these coefficient matrices for their historical achievement profiles, which in turn provide percentile growth projections/trajectories that indicate how quickly improvement must take place for future achievement targets to be reached.

Educators and students can access SGP results either via the Data SGP summary report or the more in-depth sgpData spreadsheet for each individual student. The latter features five years of annual, vertically scaled assessment data for every student; its first column displays their unique ID number; next come columns detailing assessment scores from 2013, 2014, 2015 and 2016.

This article’s purpose is to introduce newcomers to the SGP analysis tool and sgpData spreadsheet by providing a step-by-step tutorial on their use. Although not covering every possible scenario, this tutorial serves as a handy reference guide for the most commonly utilized parts of these tools.

Although these tools and functions were developed for use with LONG data formats, they may be modified for use with WIDE formats as well. To facilitate this flexibility, the SGP tool comes packaged with an exemplar state-specific data set known as sgpData_long that serves as an example of wide analysis when working with wide datasets; its purpose is to show how SGP analysis is carried out when working with wide data sets containing multiple variables that require lower level SGP functions to operate efficiently; additionally it contains an anonymized lookup table which associates each long data record with its teacher in accordance with each long record sgpData_INSTRUCTOR_NUMBER data set which makes available download from SGP website.