Data SGP – Calculating Student Growth Percentiles (SGP) and Percentile Growth Projections (Projects)
The data sgp package contains classes, functions and data for calculating student growth percentiles (SGP) and percentile growth projections/trajectories using large scale, longitudinal education assessment data. SGP analysis uses techniques such as quantile regression and provides teachers with the information they need to plan for future achievement targets.
SGPs describe a student’s performance in a subject-matter test, such as the MCAS, relative to students with comparable score histories. For example, a SGP of 50 indicates that the student’s score on the current test is equal to or better than half of the scores of students with similar score histories on the subject-matter test who demonstrated growth at the same rate.
An SGP is calculated by comparing up to two years of MCAS test results. For each year of testing, a student’s score is compared to the scores of academic peers in their grade who have taken MCAS tests at the same time. Academic peers are matched on several dimensions, including demographics such as gender, income and educational programs such as sheltered English immersion or special education.
The SGP vignette below explains how these comparisons are made. The class SGPdata provides 4 exemplar data sets that can be used for SGP analyses. One, sgpData_LONG, specifies data in the wide format used by lower level SGP functions like studentGrowthPercentiles and studentGrowthProjections. The other two, sgpData_INSTRUCTOR_NUMBER and prepareSGP, provide teacher-student lookup tables that are utilized by higher level SGP functions like abcSGP, analyzeSGP, and prepareSGP.
While these exemplars are useful for understanding the processes behind SGP calculations, they should not be interpreted as the “typical” growth seen by students in a given year. This is because the SGP distribution across the state is expected to follow a bell-shape curve, with roughly an equal number of students at each percentage point (10 or decile).
As such, the mean SGP reported for a school/district or student group is not indicative of the average amount of growth that a group of students would typically experience. However, it is important to note that in some years, a small proportion of students may demonstrate more or less growth than others in the state. This may be caused by a variety of reasons such as the Covid-19 pandemic, the economy or other factors beyond the control of individual schools and districts.
For this reason, it is important to consider the context in which a group’s SGP is being reported when making decisions about educator evaluation. SGPs are only released to teachers for high-stakes use when they have been analyzed at the group level and are deemed to be reliable and valid. SGPs cannot be trusted at the school/district or group level for educator evaluation purposes until high-stakes use is implemented in 2018. The following vignettes describe how to interpret the average SGP of each of the groups displayed above. More information about the calculation and interpretation of these SGPs can also be found on MDE’s website.