Getting Started With the Data SGP Package
The sgp package provides classes, functions and data for calculating student growth percentiles and percentile growth projections/trajectories using large scale, longitudinal education assessment data. Percentile growth plots illustrate a student’s progress over time and provide information about what future achievement targets are expected or required for the student to reach. Percentile growth estimates are computed by using quantile regression to estimate the conditional density associated with a student’s prior test scores. This is then used to calculate the predicted score needed to achieve a future test score target (or, alternatively, the percentile of a trajectory).
A student’s growth plot shows their progress over time in terms that are familiar to teachers and parents, the percentage. This is because the growth plot is based on comparisons to students with similar prior test scores (a student’s academic peers). The plot can be sorted by the percentage of the student’s score that remains at each point in time and the corresponding graph can be used to determine how much a student has to grow in order to achieve their next test score target.
Getting started with the SGP package is fairly straightforward but running analyses does require some familiarity with the R software environment. The software is available free on most operating systems including Windows, OSX and Linux. There are many resources for learning how to use R on the CRAN website.
To get started, the user needs to load the data sgp package and a set of analysis functions. The sgpData set of data is an anonymized, panel data set consisting of 5 years of annual, vertically scaled, assessment data. The first column in sgpData provides the unique student identifier and the next five columns provide the grade level associated with the students assessment occurrences.
The final column provides the scaled assessment score associated with the students grade level. This is the number that will be used in the studentGrowthPercentiles and studentGrowthProjections functions to produce percentile growth estimates and predictions.
The sgpData_INSTRUCTOR_NUMBER table is an anonymized, teacher-student lookup table that provides the insturctor numbers associated with each of the student’s test records. If the sgpData_INSTRUCTOR_NUMBER column is present in a student’s test record then a corresponding teacher-student-trajectory graph will be produced for that student.