The data sgp website is an innovative new resource designed to assist educators and stakeholders interpret Student Growth Percentiles (SGP). It provides an overview of how SGPs are calculated as well as steps for downloading individual students’ SGP data. Furthermore, podcasts help users navigate this tool while understanding its data.

SGPs are calculated by comparing a student’s assessment score for this year against their performance on two previous years’ assessments in the same subject area. This allows educators to easily identify growth for students who typically achieve high scores but whose performances vary year to year.

While this is an effective first step, it should be remembered that differences between years of SGPs should be interpreted with care as they are calculated from scratch each year. Furthermore, differences across content areas must also be treated with caution as SGPs take into account students’ experiences since 8th grade graduation.

This site allows for the visualization of student growth percentiles at either the school/district or student group levels with an option to drill into individual student data points that includes multiple data points. Authorized users may view each student’s SGPs for both this year and previous two years; academic peer groups including all those scoring similarly on MCAS exams; individual growth percentiles (high, low or typical); scale score data as well as attendance rates in their growth year.

SGPs for each student are determined using quantile regression, a statistical method which compares assessment scores against those of academic peers across the state. This process considers factors like demographics, educational programs (e.g. sheltered English immersion and special education), achievement levels and more when making recommendations.

SGPs for all students are released annually during springtime on the BAA Secure Site, starting with SGPs for grades 4-8 in 2008-09 school year and moving through grade 9-11 starting in 2017-18 school year.

As with any analysis, SGPs’ success depends heavily on data preparation. The lower level functions that perform calculations – studentGrowthPercentiles and studentGrowthProjections – require WIDE formatted data. Meanwhile, higher level functions abcSGP and updateSGP provide wrappers around these functions to “simplify” the source code associated with analyses by consolidating steps into single functions.