Data sgp is an analytical tool to assist school districts and individual students with understanding growth models for state assessment data. Its use is primarily driven by the need to compare student performance within the context of the overall test administration and to determine whether the student’s current achievement is above, below, or at expected levels. This analysis can be conducted at the student, school/district, or subgroup level and is typically a part of teacher evaluation processes.
To conduct SGP analyses one needs a computer running the open source statistical software R. This software is available for Windows, Mac OSX and Linux and can be downloaded from CRAN. The bulk of the work in conducting SGP analyses is the preparation of data. There are two primary steps to the analyses: data preparation and data analysis.
The sgp package, which can be accessed through the RStudio interface, provides tools and functions to prepare longitudinal (time dependent) student assessment data for SGP analysis. The package also includes exemplar WIDE and LONG format data sets to assist with this process.
An important point to note is that a student’s growth percentiles are based on their relative performance compared to academic peers with similar score histories on MCAS tests administered in previous years. The academic peer groups are determined using a process called quantile regression, which places the scores of students with similar score histories on a normative scale. These percentile ranks are then used to identify the relative performance of a student’s current MCAS score.
In addition to the academic peer groups, a student’s growth percentiles can be impacted by various factors such as demographic characteristics (e.g., gender, racial/ethnic identity, educational programs like sheltered English immersion or special education) and student achievement levels from other assessments such as district or local tests. For this reason, a student’s growth may appear different across grades, even for the same subject area.
For example, a student who scored a high number on the 2024 MCAS English Language Arts (ELA) test will have higher SGPs in that subject than a student with a lower score. This is because the higher score is more likely to indicate high student growth.
For these reasons, it is critical that the SGP results be viewed in context and not by themselves. It is also important that the SGP results be interpreted in conjunction with other measures of student achievement and progress, including achievement gaps and proficiency rates. This article describes the various SGP analyses that can be run and provides tips for interpreting the results. We hope you find it helpful! Please contact us with questions or suggestions for future articles.