Posted on

Data SGP – How to Analyze MCAS Student Growth Percentiles

Student Growth Percentiles are the standard measurement of students’ performance on statewide assessments. They provide a more accurate picture of individual student performance than raw scores alone, which can be misleading when comparing the performance of different students with identical starting points. For example, Student A scored a 300 on this year’s MCAS English language arts (ELA) exam, while Student B scored a 700. A simple comparison would suggest that Student A had higher scale score growth than Student B, but this is not the case since the two students had very different MCAS scaled score histories. A more accurate method of evaluating student growth is to compare the student’s current SGP to the SGP of their academic peers.

Student growth percentiles are calculated using up to two years of historical MCAS data. SGPs are based on a statistical procedure called quantile regression that compares a student’s performance to the performances of academic peers who scored similarly on previous MCAS tests. These academic peers are identified statewide by grades and content areas, and include students from all demographic groups including special populations (e.g., low income and sheltered English immersion).

A student’s SGP indicates how much more their performance improved from one year to the next relative to their academic peers. A high SGP means that the student’s performance improved more than most of their academic peers, while a low SGP means the student’s performance declined more than most of their academic peers.

While SGPs are being released in order to familiarize educators with the new measures, the MDE has made it clear that they will not be used for educator evaluations until 2018/19 and beyond. This allows the SGPs to stabilize for three years before they can be used for high stakes decisions.

SGP analyses are complex and time consuming, and the bulk of the work is in preparation and analysis of the data. We are here to help schools and districts make the process as straightforward and manageable as possible.

The data sgp package contains 4 example data sets for conducting SGP analyses. The first, sgpData, specifies data in the WIDE format that’s used by lower level SGP functions like studentGrowthPercentiles and studentGrowthProjections. The other two data sets, sgpData_LONG and sgptData_LONG, specify data in the LONG format that’s used by higher level SGP functions such as prepareSGP and analyzeSGP. For all but the simplest, one off, analyses you will likely be better off using the LONG data set formats as they have many preparation and storage advantages over WIDE.

The data sgp package also includes an R script for generating SGP lookup tables which can be downloaded to school information systems to permit additional analyses. SGP lookup tables are used to generate student growth projections and trajectories, and to help interpret the results of other SGP analyses. This script is available from the GitHub site and can be run from a command line. More information about the script can be found in our documentation.