The data sgp package provides classes and functions to calculate student growth percentiles and percentile growth projections/trajectories using large scale, longitudinal education assessment data. The package utilizes a combination of quantile regression and derived coefficient matrices to estimate the conditional density associated with each students achievement history. The percentile growth estimates are then used to generate projections/trajectories that show the percentile increase needed to reach future achievement targets.
To make use of the functions within this package, users must be familiar with the R programming language. This is an open-source programming language that is available free of charge for Windows, Mac OSX or Linux. While many of the SGP analysis functions are straightforward, there are a number that require more advanced knowledge of R. If you are new to R, there are numerous resources on the CRAN website that can help you get started.
Getting Started
To use the data sgp package, you must first install R on your computer. This is a free, open-source software program that can be downloaded from the CRAN website. Once installed, you can access the SGP functions by entering sgp into your command window. The data sgp package is also compatible with Python, which allows you to integrate the tools into your existing workflows.
Before starting the SGP analyses, it is important to understand the structure of your data set. For example, the SGP package requires an anonymized, panel data set with 5 years of annual, vertically scaled assessments in WIDE format. The sgpData data set provided with the SGP package demonstrates how this format should be structured.
SGP is a method of assessing students’ performance based on their progress over time, relative to other students with similar prior test scores (their academic peers). Student growth percentiles are calculated from this data and provide a more accurate representation of a student’s true achievement levels than raw score percentiles alone.
Although student growth percentiles are a useful metric for understanding a student’s progress, they should be used in conjunction with other measures of achievement. For example, student growth percentiles can be compared to the average of the students’ previous test sections. In addition, students with very high raw score on previous tests can still have low SGP because their previous performances were so strong.
For this reason, it is recommended that you always analyze your data with both the SGP and the raw score percentageiles. In most cases, this will be the best way to ensure that your results are valid. The raw score percentageiles are the more traditional measure, and they should be considered when comparing student results to other schools in your state. The SGP percentageiles, on the other hand, are more appropriate for evaluating individual student growth in a particular subject area. For this reason, they are most commonly used to compare the results of gifted and talented programs.