The data sgp package provides users with a set of functions for performing Student Growth Percentile and Student Growth Projection analyses. It requires the use of the R software environment which is available for Windows, OSX and Linux. It also requires familiarity with the basic concepts of student growth models. A good starting place is the article A Practitioner’s Guide to Student Growth Models by Damian W. Betenbenner and Adam R. Van Iwaarden.

Data sgp is an aggregation of standardized test scores from multiple assessments and years. It is designed to provide a summary of student performance, including normative and non-normative growth patterns. This information can be useful in identifying the areas where students need to improve, as well as in making comparisons across schools and grade levels.

Unlike standard assessment reports which indicate a student’s achievement, Student Growth Percentiles are relative measures of student performance that allow comparisons between students across content areas and over time. Student Growth Percentiles are calculated using a statistical methodology that compares students to the highest performing student in their cohort. Hence, a student with a higher Student Growth Percentile than another student is not necessarily more proficient in a particular area of the curriculum.

In order to calculate student growth percentiles and projections, the data sgp package uses a statistical method called normative growth analysis. This method compares student performance to the highest performing student in their cohort, and then estimates how far each student should grow in order to achieve this level of proficiency.

The function sgptData_LONG contains an anonymized panel dataset of 8 windows (3 windows annually) of student assessment data in long format for 3 content areas. This data is accompanied by teacher identifiers which can be used to create an instructor lookup table for each student’s test record. A variety of student categorization variables are included as well as demographic variables for creating school/district aggregates.

The sgp_percentiles and sgp_projections functions are wrappers for lower level SGP calculations. These functions are typically run with LONG formatted data as they utilize the embedded state related meta-data (e.g. knot and boundary, cutscores, CSEMs) that are stored in the SGPstateData object.

In general, it is recommended that users prepare their data sets in the LONG format rather than WIDE as most of the higher level SGP analyses require this format. Any errors that arise when running these analyses generally revert back to data preparation problems so it is best to start with properly prepared data. This will save time and effort in resolving issues that may arise during the actual analysis process. If data preparation is done correctly, most SGP analyses are very straightforward to run. This makes it a very quick and easy way to access high quality student growth information.