Getting Started With Data SGP
Beginning Data SGP Since Data sgp utilizes open source R software for student growth percentile analyses, its use requires some familiarity with its environment. However, most time spent with SGP analyses focuses on data preparation; once complete, running SGP calculations becomes straightforward.
The data sgp package offers classes, functions and data that enable users to calculate student growth percentiles (SGP), their projections/trajectories and conditional density matrices for every student in a system. It employs large scale longitudinal assessment data combined with quantile regression techniques in order to accurately calculate each individual student’s conditional density matrix over their achievement history.
SGPs are calculated based on the percentage of students with similar score histories as one being assessed and provide insight into how much growth a given student needs in order to reach his or her desired percentile ranking in a subject-matter test. An SGP value of 50 indicates that this student has demonstrated growth equivalent to or exceeding half their academic peer group in one subject matter test subject matter test and are reported at their nearest decile.
As part of your interpretation of SGPs, it is essential to keep in mind that percentile rankings are calculated annually and thus differences between years should not be taken at face value. As seen below, two students who scored identically on an MCAS exam in 2023 both earned an SGP ranking of 50 but Student A is closer to the top academic peer group as indicated by their relative SGP ranking.
As well as calculating SGPs, this package offers other reports such as teacher-student reports and school/district/student group average SGPs as well as graphs showing student growth over time – especially useful when assessing educator performance.
Data SGP Package Provides Example Longitudinal Assessment Data Set The data sgp package offers an example data set (sgpData), that serves as an illustration for setting up longitudinal assessment data in WIDE format – each row represents an individual student while columns contain variables associated with them at various points in time. It also offers LONG formatted sgpData_LONG that will assist with setting up longitudinal data using SGP.
As opposed to using either of the sgpData_LONG and sgpData_WIDE data sets for SGP analyses, sgpData is recommended because it makes accessing teacher-student lookup information necessary for projection/trajectories and conditional density matrices much simpler. For more details regarding setting up data for an SGP analysis please see the Data SGP Package Documentation.