Data sgp leverages longitudinal student assessment data to produce statistical growth percentiles (SGP) for individual students, groups of students, schools and districts. This information can then be shared with educators and parents in terms that are familiar to them – percentile ranks of how much relative progress a student has made compared to academic peers who started out at the same starting point. This information can be used to inform classroom practices, evaluate teachers/schools/districts and support broader research initiatives.
SGPs are calculated on a 1-99 scale, with lower numbers indicating less relative student growth and higher ones indicating more. They are determined by measuring the magnitude of a change in a student’s raw test score compared to the average of the group (academic peers with similar prior assessment scores). SGPs are produced for each of a student’s assessments, and may be reported as “Window Specific SGP” or “Current SGP.”
Both of these types of SGPs are important to report, but it is essential that educators consider both when making educational decisions for their students. SGPs should be considered alongside a number of other metrics including classroom observations, teacher evaluations and standardized test scores.
Running SGP analyses requires access to longitudinal student assessment data in WIDE or LONG format. The bulk of the time that is spent on SGP analysis is in the preparation of this data and most errors associated with running these analyses are a result of improper or incomplete data preparation.
The SGP vignette that is available on the sgp website provides an overview of the necessary steps for preparing this type of data for analysis with the sgp tool. This vignette also includes a spreadsheet, sgpData, that is formatted for SGP analysis and provides examples of how to create the required variables for each of the SGP analyses. The sgpData spreadsheet contains both a wide and long version of the sgpData dataset (sgptData_LONG and sgpData_WIDE). The sgpData_LONG dataset includes the required variables for SGP analyses; sgpData_INSTRUCTOR_NUMBER is an anonymized lookup table that associates a teacher to each of the student’s test records in the LONG data set.
The SGP tool uses the open source programming language R, which is free-of-charge for Windows, OSX and Linux computers. Several resources on the CRAN website provide assistance to newcomers looking to learn about and use this software. The sgp tool also utilizes some advanced R functions that require some level of familiarity with these capabilities, and the sgpData vignette includes an explanation of these functions and their usage. The lower level function studentGrowthPercentiles and the higher level wrapper function studentGrowthProjections are designed to be used with either a wide or long format data set and utilize the sgpData vignette for further guidance on creating these types of data sets.