The data sgp package is an open-source software tool that allows for the calculation of student growth percentiles and projections/trajectories using large scale, longitudinal education assessment data. The data can come from standardized tests, portfolios, grading scales and/or other sources. Using these tools, educators can identify students who are at risk of not meeting academic potential and provide them with additional support. They can also evaluate current educational systems and policies to determine how they can be improved.
Data SGP uses the historical growth trajectories of Star examinees to map out what a students potential future performance could look like, including the amount of necessary growth for a student to maintain proficiency and beyond. Using this information, educators can predict what a students performance will be on the next examination and can use that information to help guide their instruction. This is known as a projected score. This information is incorporated into the student’s profile on the SGP website and is available in the Star Growth Report for each individual student by selecting a past or current window.
This method of predicting student performance is more accurate than standard test scores and can capture the complex nature of learning. In addition, it does not rely on the assumption that all students are equally proficient and can be used to measure student achievement with a variety of methodologies. As such, it is an ideal candidate for a new accountability system that focuses on student growth rather than simply test score performance.
To run an SGP analysis, the user must have access to a dataset of student assessments in the LONG format, along with any state level student aggregates provided by the sgpstateData metadata-data. These analyses can be complicated and time consuming, so the proper preparation of data is critical for successful results. If errors are encountered, they typically revert back to data preparation issues so it is important to have the appropriate tools/hardware and know how to prepare the data correctly.
The minimum dataset required is sgptData_LONG, which provides 8 windows (3 annually) of assessment data in the LONG format from the Early Literacy, Math and Reading content areas; a state level student aggregate is also required as provided by sgpstateData. The SGP package provides lower-level functions, such as studentGrowthPercentiles and studentGrowthProjections that are designed to work with WIDE formatted data sets and many of the higher-level function provide wrappers for sgpData. Please see the SGP Data Analysis Vignette for more comprehensive documentation on using sgpData and other wide-format data with the SGP package.
The first column of the sgpData file provides each unique student identifier, while the second through fifth columns contain the grade levels associated with the student’s assessment scores for each year (2013, 2014, 2015, 2016 and 2017). The sixth column, sgpData_INSTRUCTOR_NUMBER, is an anonymized, student-instructor lookup table that provides insturctor information associated with a students test record for a given window. This is a requirement for the generation of projected scores in both the SGP and the Star Growth Report.