Data SGP (sgp) is an open source software package that offers students, teachers and school leaders with analytical tools for school performance evaluation and improvement. Run by R software environment on Windows OSX and Linux computers alike; data sgp supports graphic analysis as well as statistical modeling and predictive analytics through its open nature.
Students are organized according to academic peers who scored similarly on past MCAS tests (their academic peer group). A statistical procedure called quantile regression places scores on a normative scale — in other words, their percentile rank within normalized score distribution for that subject area — which allows us to quickly and efficiently identify students’ relative performances both within their grade as well as across schools, districts and the state as a whole.
Student growth percentiles provide us with valuable insight into whether a student is progressing or stagnating, particularly among low performers who may need extra assistance to show how far along they have come in terms of progressing academically. Their percentile rank provides us with insight into this progressing or not.
The data sgp package offers several functions to assist with creating and analyzing student growth percentiles. studentGrowthPercentiles() takes in information from teachers for each student in an MCAS assessment and creates individual student growth percentiles for all content areas evaluated by this exam.
StudentGrowthProjections() allows for the generation of projections across multiple time frames for students. In addition, currentSGP() provides an update of a student’s most recent growth percentile across their assessments to track progress over time.
Although a student’s SGP can vary year to year, its average student growth percentile remains constant and indicates an overall pattern of student learning. It is important to keep in mind that their average SGP reflects their experience across two years – any interpretation should take this into account when making conclusions from this figure.
Students with high MCAS scores tend to experience more volatile student growth percentages (SGPs). This variation could be the result of both inherent test variability, as well as various levels of learning over the years for these students. Regardless, the statewide median SGP should remain at about 50.
The sgpData dataset features anonymized student-instructor lookup tables to provide instructor details linked to students test records. This is necessary since any given test record could have multiple instructors associated with it; additionally, any given instructor could have multiple students associated with their test record. The sgpData_INSTRUCTOR_NUMBER table serves to standardize this presentation of information while providing schools an exemplar set of test records that demonstrate what data should be provided when uploading their own test records to sgpData.
