Whether it’s for students in kindergarten or the top of high school, the success of any education program depends on its ability to improve the knowledge, skills and abilities of all students. To do this, programs need to measure student growth in ways that are reliable and valid – not just for the bottom half of the class, but also for the top half. To do this, they need data sgp.

Unlike traditional test scores, which are calculated for groups of students of comparable starting points, SGPs are based on the individual histories of each student. This means that the results of a single assessment can be used to estimate what a student would achieve on an upcoming assessment, given their previous history and current performance. This approach to measurement enables educators to identify students who need additional support and help, as well as to differentiate instruction based on student need.

While SGPs can be analyzed at many levels, the most important is at the classroom level. In a classroom, SGPs can be used to determine the appropriate placement of students in academic courses or to determine if a student is ready to graduate. In addition, SGPs can be analyzed to evaluate the effectiveness of a teacher or school. The simplest form of SGP analysis involves comparing the progress of a student to that of students in the same grade level at their school who are taught by a similarly qualified educator.

SGPs can be analyzed at the student, school and district levels to understand how different educational programs are performing and what they need to do to improve student achievement. This requires accurate and accessible data that includes historical performance, growth trajectories and projections. It also requires a clear and transparent communication of these findings to all stakeholders.

Fortunately, the sgpdata package provides an exemplar WIDE and LONG formatted data set to facilitate this work. The sgpData_WIDE dataset models the data structure required to support the lower level functions studentGrowthPercentiles and studentGrowthProjections while the sgpData_LONG dataset represents longitudinal (time dependent) data for each individual student.

Both the sgpData_WIDE and sgpData_LONG datasets include an anonymized, teacher-student lookup table that allows students to be associated with multiple teachers in a content area for a given year. This is a necessary feature of SGP analysis, as relationships between true SGPs and student background covariates are likely to depend on PTh, and the distributional properties of this function can only be assessed with the availability of full longitudinal data.

As SGP analyses become increasingly operationally driven, it is likely that many states will require the use of LONG formatted data. Most higher level functions in the sgpdata package are designed to work with long data and assume that it is managed via the embedded sgpData stateData meta-data. This is preferable to using WIDE data if updating analyses for each new assessment year because of the simpler management and preparation of long data sets.