Data SGP and Its Limitations

Data SGP (Student Growth Progress) is a statistical measure of student learning and progress calculated using longitudinal data. While educators can use it as one metric when making decisions and teaching practices, it should not be the sole determinant of student success; educators must evaluate students using multiple measures – including test results and classroom performance evaluations. Furthermore, educators should understand its limitations to avoid false or misleading conclusions being drawn from Data sGP measurements.

Student Growth Percentiles (SGPs) provide an individual rank within a cohort from 1-99 which indicates each student’s improvement or decline over time. A high SGP indicates more impressive gains while a low one may signify struggling or no progress being made at all. An SGP is determined by comparing individual assessment scores on an exam section with scores from others in their cohort with similar academic histories, then weighting it for more accurate ranking of student growth.

SGPs can be an invaluable educational tool for both teachers and parents, but should never be seen as the sole indicator of student success or failure. When making instructional and curricular decisions based on SGP data alone, educators should take into account additional data points like classroom teacher evaluations and student performance in class; SGPs may produce spurious correlations that are caused by differences between school/teacher characteristics or baseline cohort design, leading to false positive correlations that may be misleading for those students who already possess exceptional academic abilities.

R’s SGP package provides tools for calculating, displaying and reporting on SGPs. Most of its lower level functions, including studentGrowthPercentiles and studentGrowthProjections, require wide formatted data; please see the SGP data analysis vignette for information on how to create such data sets. Furthermore, this package also offers two LONG formatted datasets which allow users to conduct SGP analyses using LONG formatted data sets sgptData_LONG and sgpData_LONG_INSTRUCTOR_NUMBER which enable users to perform analyses using LONG formatted datasets.

To obtain SGPs, users must have access to longitudinal student assessment data in the appropriate formats for each year of testing. The sgpData data set contains assessment records in wide and long formats for all five years of testing; its associated Instructor Number Lookup Table can also be accessed.