Data sgp provides educators with a tool to understand student growth percentiles. A student’s SGP indicates how much he or she has progressed relative to peers with similar MCAS performance histories. SGPs range from 1 to 99 and are interpreted like percentile ranks, where higher numbers indicate greater relative growth.
Student growth percentiles are calculated based on the most recent test score and one or more previous test scores from a previous testing window. For example, a student’s SGP for the 2015-16 school year may be based on his or her most recent MCAS score and one or more prior tests from the Forward Exam or WKCE testing windows. While incorporating multiple years of test score history is essential to generating reliable and valid student-level growth models, this approach introduces additional complexity and uncertainty into the model results. As a result, students’ SGPs can be more vulnerable to spurious correlations with classroom composition or the design of a baseline cohort than VAMs.
SGPs can be used in combination with achievement targets and goals to provide a clearer picture of what it takes for students to succeed. The goal is to identify high achieving students as well as those who need additional support. SGPs can also help educators see how their current instruction is assisting students to meet their achievement goals.
The sgpData spreadsheet, which is part of the data sgp package, makes it easy for educators to compare their students against peers and across different grades and subject areas. The spreadsheet also includes information about a student that is not available in the SGP summary report, such as their gender and socioeconomic status.
Educators can run a series of SGP analyses in the R statistical software environment with the data sgp package. Data sgp requires a computer that has the free and open source program R installed. R is available for Windows, OSX and Linux and is simple enough that it can be learned quickly. A basic familiarity with R is sufficient for running most SGP analyses.
To perform SGP analyses, an educator needs to prepare a dataset that includes sgpData and an instructor-student lookup table, sgpData_INSTRUCTOR_NUMBER. The analysis functions in the sgpData package include wrappers for the lower level calculations, studentGrowthPercentiles and studentGrowthProjections. These wrappers require LONG formatted data. It is recommended that LONG formatted data be used for operational analyses as most of the package capability is designed around it.
Almost all errors that occur during an operational SGP analysis revert back to problems in data preparation. Therefore, it is important to follow the steps in this vignette and to make sure that your data is set up correctly before attempting any analysis. After a thorough check, SGP analyses are straightforward and can be completed in less than an hour. If you encounter any issues while performing these analyses, please let us know by submitting an issue on GitHub. This will enable us to better track and prioritize fixes to the sgpData package.