The data sgp package is an easy-to-use Python library for conducting SGP analyses. It enables teachers and administrators to quickly identify whether students grew more or less than their academic peers over the course of several assessment windows. It can also be used to determine how well students are progressing toward their growth targets.
In order to conduct SGP analyses using the data sgp package, one must first prepare the data sets that will be used for each analysis. This is a simple step that can be done by either using the provided exemplar data sets or by writing one’s own data set that meets the specifications for calculating SGP percentiles and projections. Once the data sets are prepared, the user can create SGP objects that will display both window specific SGP percentiles and current SGP scores for a given student.
SGPs are based on a percentile ranking of all students who have taken the same assessments. This means that when interpreting SGPs, it is important to remember that SGPs are calculated anew each year and that differences in SGPs between years should be interpreted cautiously (i.e., a higher SGP score in 2023 does not necessarily indicate greater relative growth over the course of the school year than a lower SGP score in 2022).
The sgptData_LONG exemplar data set provides 8 windows (3 windows annually) of assessment data in LONG format for three content areas. It contains demographic variables that are required when running SGP analyses, as well as student categorization variables needed to create student aggregates by the summarizeSGP function. The sgpData_INSTRUCTOR_NUMBER is an anonymized, teacher-student lookup table that identifies the instructor associated with each student’s assessment record. The data sgptData_LONG and sgpData_INSTRUCTOR_NUMBER are necessary when using the summariesSGP function to produce window specific SGP percentiles or the summariesSGP function to generate projected growth percentages.
The data sgp package has wrapper functions called abcSGP and updateSGP that simplify the steps of creating a SGP object for use in operational analyses. These functions take the exemplar data sets and student-instructor lookup table and create an SGP object that can be used to run the lower level studentGrowthPercentiles and studentGrowthProjections functions. Both of these SGP objects can be saved in order to be re-used for future analyses. The data sgp package also has other functions that simplify common tasks that are frequently conducted in operational SGP analyses. These include creating a plot of a single student’s SGP percentiles over time and creating student growth and achievement plots for a group of students. The full list of available functions can be found here. Please contact us if you have questions about the data sgp package. We are always happy to help. We are also open to feature requests and contributions. You can leave a comment here or submit an issue on GitHub. We are a small, dedicated team of education professionals working together to make the best possible tools for our users.
