Data SGP is an analysis tool for longitudinal (time dependent) student assessment data to create statistical growth plots which show the relative performance of a student over time compared to his or her academic peers. This information can be useful in assessing and informing instruction, educator evaluation systems and supporting the state achievement target/goals used for accountability. It offers a more accurate measurement of growth than traditional percentile scores do.

SGP analyzes standardized assessment test data to create growth charts for each individual student. The data is then aggregated to report student/teacher performance on a scale of 1-99. The higher the score the more a student has shown growth. A score of 75 or above indicates a student has performed better than the bottom 75% of his or her academic peers over time.

While the term ‘big data’ has become somewhat of a buzzword in modern education, this type of research and analysis represents relatively small potatoes by comparison to the vast amount of information available for analytics on global Facebook interactions or the amount of medical records that can be analyzed for predictive health care. However, it is still a tremendous leap in computational capabilities for educators and researchers and requires some new learning to effectively manage.

The data sgp package contains tools for conducting these analyses, from creating and managing the data to creating and reporting the results. It also provides higher level functions that “wrap” the lower level studentGrowthPercentiles and studentGrowthProjections to simplify the source code necessary for operational SGP analyses.

A key feature of SGP that separates it from other methods is the ability to compare student/teacher performance against state achievement targets/goals. This can be a powerful motivating force for teachers to meet or exceed official goals/targets and a useful measure for schools/districts of effectiveness.

SGP utilizes latent achievement trait models estimated from teacher evaluation criteria alongside historical test score histories to establish growth standards that minimize estimation error and improve validity when comparing students over time. This method is different from value-added scores which are based on a more complicated model of student/teacher performance that can introduce significant measurement error.

The sgpdata package provides an example WIDE format data set (sgpData_WIDE) and a LONG format file (sgpData_LONG) to simulate the time dependent data used with lower level SGP functions like studentGrowthPercentiles and studentsGrowthProjections. The package also includes an exemplar SGPdata lookup file sgpData_INSTRUCTOR_NUMBER to facilitate converting this data into the required SGPdata format for higher level functions. This is the recommended format for operational analyses.