Student Growth Percentiles (SGPs) measure how well current assessment scores measure up against academic peers based on performance on previous assessments. An SGP of a student serves as an indication of whether their performance meets or surpasses state expectations; they’re calculated using quantile regression models and data from prior years to predict scores on current tests.

Student Growth Profiles (SGPs) measure academic achievement at various grade levels. While those with low SGPs demonstrate below-average achievement and may need extra support in making progress. SGPs help educators to identify these students, identify learning opportunities necessary, and inform educator evaluators about what may occur this year should their student continue on the same trajectory or accelerate their growth.

SGP analyses are performed in the R software environment, which is free and available for Windows, OSX, and Linux computers. SGP analyses require at least 4GB of memory in your system to run effectively; additionally the sgp package offers two wrapper functions called abcSGP and updateSGP which simplify lower level SGP functions studentGrowthPercentiles and studentGrowthProjections while decreasing source code requirements – they should be recommended for operational use.

SGP calculations can be complex and involve many steps, often leading to errors due to improper or incomplete data preparation. We advise districts using the BAA secure site tool – data SGP Tool – for their analyses and preparation needs.

There are a few basic guidelines to be kept in mind when using data sgp for the most frequent student and district-wide SGP analyses. First, SGP analysis tools require long-form LONG data that contains seven variables including: VALID_CASE, CONTENT_AREA, YEAR, ID, SCALE_SCORE and ACHIEVEMENT_LEVEL – this data set also contains the teacher lookup file named sgptData_INSTRUCTOR_NUMBER for convenience.

Students must enroll in courses for which there is valid MCAS test data available and where a teacher has been designated as “teacher of record.” The sgptData_LONG dataset offers English Language Arts and mathematics examples to support this configuration.

One key distinction between the sgpTData_LONG and sgptData_LONG files is that sgptData_LONG also includes demographic variables for each student that can be used to generate reports such as teacher aggregates; by contrast, sgpTData_LONG does not.