OSPI Student Growth Percentiles
Student growth percentiles (SGP) are a powerful tool for sharing academic information about student achievement in terms familiar to teachers and parents – percentages. They are calculated by ranking a student’s change in scale score relative to the change of a group of students with similar starting scores (their academic peers). A student’s SGP rank is displayed on a scale from 1 to 99, with higher ranks indicating better performance.
SGPs are useful for educators because they provide meaningful information about a student’s learning gains based on comparisons to similar students. This information can help teachers identify areas for improvement and make decisions about what instruction is most appropriate for each student. The tool is designed to be as user friendly as possible, and OSPI has a number of resources available for training on the use of SGPs.
The SGP data set contains up to five years of student assessment score data and is organized into two tables: sgpData and sgpData_LONG. The sgpData table contains student information associated with the same test record at each point in time, while the sgpData_LONG data set includes additional variables for longitudinal analysis such as a student’s grade level and their class. The SGP package offers both wide and long format data sets for analyses; however, most users will likely find it easier to use the long data format sgpData and the higher level SGP functions studentGrowthPercentiles and studentGrowthProjections. For more information on how to work with these data, see the SGP data analysis vignette.
Using SGP data with the R programming language requires some computer proficiency. Many users will benefit from familiarity with the basics of the R environment. There are a number of resources for learning about R, including online tutorials and a free, open-source programming language distribution for Windows, Mac OSX or Linux. The SGP package is built on the CRAN version of R and uses advanced functions that are not suitable for basic usage.
Aside from software and computer requirements, the bulk of the time required for an SGP analysis is spent preparing the data for the analysis. We recommend following the SGP data preparation vignette to ensure that the resulting analysis will be accurate and reproducible.
While SGPs are valuable tools for sharing student performance information, they can be misleading if used improperly. SGPs must be carefully evaluated for the validity of the comparisons made and for potential distortions caused by differences in the school/district context, baseline cohort design and teaching methodologies. These distortions are typically a result of a combination of factors and may be difficult to identify. This is one reason that SGPs should only be used by trained staff.