Data SGP is a statistical analysis package that provides a suite of functions for generating student growth percentiles and growth projections/trajectories using large scale, longitudinal education assessment data. The package uses a quantile regression methodology to establish conditional density models of achievement and then derives student growth projections from these estimates.
SGP allows users to compare the performance of students across grade levels and between school districts or schools within a district. The growth percentiles provide a way to see how well students are progressing towards their goals and identify students that need intervention and support. SGPs are calculated by comparing a student’s score on a state assessment to the scores of their academic peers who took the same assessment. This means that a student who has a high score can be expected to have high growth while a student who scored lower can expect low growth.
For example, a student who has a 3rd grade SGP of 85 shows that she has performed better than 85 percent of her academic peers in the same grade. However, if a student has an SGP of 99 that indicates she has not made any progress towards her goal for the year.
Students who score at the highest end of the scale (e.g., the top 25 percent of students) are also expected to have high growth, but that doesn’t always happen. For this reason, the SGP model takes into account a number of other factors that contribute to student success beyond simply testing ability. For example, the model takes into account a student’s overall attendance, their participation in afterschool programs and other activities, and the amount of time spent studying.
Using the data sgp package requires an R software application that is available for Windows, Mac OSX and Linux. The software is free and available from CRAN. Running SGP analyses does require some familiarity with the R programming language but there are many resources available to help get started.
The SGP package makes use of the wide-format data format sgpData which contains information about all students who have taken state assessments in each of the past 5 years. The first row of the data file contains the ID that uniquely identifies each student. The following five rows provide the assessment scores for each of the students in the sample. The last row, SS_2013, SS_2014, SS_2015, SS_2016, and SS_2017, provides the scale scores associated with each of the assessments in those years.
The SGP package requires a quad core machine with at least 4 GB of memory. Running the SGP demonstration on a large set of data can take as long as several hours. It is recommended that you run the demonstration on a dedicated machine for optimal performance. For more detailed explanations of the SGP analysis process and its various methods see the vignette that accompanies this article.