Student Growth Percentiles (SGPs) have become an innovative trend in data analytics, providing school leaders with a valuable tool for tracking students’ performance over time and making decisions regarding students, teachers and the overall organization.
An easy way to understand SGP: Consider sixth grader Simon who earned a scale score of 370 on this year’s state English Language Arts (ELA) test and, when compared with his academic peers, showed an 85 growth percentile, which indicates 85% more progress in ELA than Simon did – this could indicate his growth towards being on track in ELA.
Note that growth percentiles are calculated using quantile regression, not traditional bell curves. They use multiple years of MCAS data and enable comparison of student performance with that of his academic peers both within their grade and across years; providing more precise measurements of a student’s achievement relative to others in his or her academic group.
SGPs are calculated for students statewide in grades 5 through 11 as well as each student’s academic peers within his grade. Academic peers are divided by socioeconomic status, race/ethnicity/gender demographic groupings and educational program (e.g. sheltered English immersion/special education). Their performances on previous MCAS assessments are then ranked according to each student’s relative performance compared with his academic peers – this enables us to determine how many of his academic peers outshone him on the latest MCAS assessment.
Student Growth Projection Analysis (SGPt), is a statistical technique which computes student assessment results over multiple years to provide a long-term performance estimate of each student. SGPt also gives an indication of current performance levels which can help identify any areas for improvement and measure how proficient a student currently is in his studies.
Data SGP allows us to gain valuable insight into how students, schools, and districts across the state are faring in terms of academic performance, which provides us with insight into which classrooms need improvement, where instruction needs improvement most urgently, as well as which students are reaching higher than average levels of success.
Users seeking to use Data SGP must first prepare and store their SGP calculations. The lowest level functions, studentGrowthPercentiles and studentGrowthProjections, require WIDE formatted data while higher level wrapper functions (including studentSGPMetrics ) use LONG formatted data. When preparing data for SGP analyses it is preferable to use LONG format as this facilitates future operational analyses over time as well as providing embedded SGPstateData meta-data availability – although for simple one off analyses WIDE formatted data may suffice.