Data SGP is an educational assessment data repository. It offers teachers and administrators tools for analyzing student performance as well as useful insights on growth percentiles – an indicator of how much students have improved relative to others with similar MCAS performance histories; higher growth percentiles indicate success.
Calculating Student Growth Projection or Trajectory. (SGP). A Student Growth Projection or Trajectory (SGP) is calculated by comparing a student’s current MCAS score with that from previous testing windows, repeated each year after reconciliation is complete. This value then serves to measure percentile growth projection or trajectory of that particular student.
To conduct SGP analyses, a computer with R software installed is needed – this free and open source program can be downloaded for Windows, OSX and Linux operating systems. While running an analysis is fairly simple, CRAN offers resources for those new to using R who may require assistance before beginning SGP analyses. It is suggested that those unfamiliar with using R should spend some time familiarizing themselves with it prior to running any analyses on SGP data sets.
As previously discussed, SGP calculations are straightforward when data preparation is conducted properly. Most errors found during data analyses can be traced back to improper preparation; that is why taking time and care in gathering your information prior to beginning any SGP analyses is so essential.
Though there are multiple approaches to analyzing educational assessment data, most analyses performed by the Department use SGP package – an open source tool developed by National Center for the Improvement of Educational Assessment – as it allows users to easily create reports and visualizations using MCAS data. It is easy to use and serves as an invaluable resource for educators.
Lower level SGP functions (studentGrowthPercentiles and studentGrowthProjections) need WIDE formatted data; however, their higher level counterparts (wrappers for lower level functions) can handle LONG formatted input data. If you plan on performing any SGP analyses, we strongly suggest storing your data in LONG format as this will allow you to take full advantage of all the capabilities provided by the package. An example can be seen with the sgptData_LONG data set that demonstrates this method of data organization. This file provides 8 years of vertically scaled student assessment data across three content areas for three content areas, anonymized so as to avoid being associated with specific schools or districts. This data set represents all three sets of variables required for SGP calculations: student ID and grade level; assessment occurrences and numeric scores associated with each assessment event. An exemplar data set serves as an illustration of this format – more comprehensive documentation is also available regarding its requirements.