The data sgp package provides an efficient and robust set of tools for creating student assessment data that is compatible with SGP analyses. Most errors associated with SGP analysis can be traced back to problems with data preparation, and so it is important that thorough and complete preparation is performed before any SGP analyses are conducted. The data sgp package includes two sample data sets in WIDE and LONG formats (sgpData and sgpData_LONG) to assist users in this preparation process.
The SGP data sgp package contains all of the functions required to prepare, analyze, and combine longitudinal student growth projections. The package is designed to work with exemplar longitudinal data sets, such as the Wisconsin Badger Exam, Forward Exam, and PSSAs. The package also includes an anonymized, student-instructor lookup table, sgpData_INSTRUCTOR_NUMBER, which allows districts to connect students with instructors through unique instructor identifiers associated with each students test record.
Preparing data for SGP requires careful consideration of the dimensions of the measurement model and the data source(s). For example, when combining data from multiple sources, it is critical to ensure that all of the data dimensions are the same size. In addition, the data must be appropriately sampled to provide unbiased and valid estimates.
In addition, it is often necessary to make adjustments for underlying biases and variation in the data source. This is done by using the SGP functions abcSGP, updateSGP, and transformSGP. These functions take a series of steps that are usually executed simultaneously in operational SGP analyses and combine them into a single function call, simplifying the source code.
A typical SGP analysis uses a series of transformations to normalize the raw data into a common measure of student growth. The most commonly used transformations are log-normalization and mean centering. These transformations are essential to ensure that the results of an SGP analysis are comparable across time and schools, and that the resulting trends are valid.
The final step in the SGP analysis process is to combine the results of analyzeSGP and transformSGP into a master longitudinal student record, Demonstration_SGP@Data. This process creates the consolidated data that is used in SGP reports and dashboards. The result is a longitudinal record of student growth that can be used to track individual student progress, identify strengths and weaknesses, and predict future performance.
Data SGP contains a wealth of information that can be used to assess student learning and performance, and to inform instructional decision making. To be effective, however, this data must be analyzed and reported in a way that is easy to interpret and understand. To this end, the SGPdata project provides a number of data visualization tools that can be used to help explain student growth and performance to educators and community stakeholders. These tools can be found in the Data Visualization Gallery and are available for use at datasgp.github.com/data-sgp/.