Data sgp leverages longitudinal student assessment data to produce statistical growth plots (SGP) of students’ relative progress compared to their academic peers. These growth percentiles are derived from the student’s previous test scores and covariates and can then be used to gauge whether or not the student is meeting an agreed upon growth target, for example 75% of their academic peers. Unfortunately, creating SGPs from standardized student test score histories involves complex calculations that often result in large estimation errors and make these growth percentiles nearly unusable for measurement purposes.
The SGP methodology provides a means to determine what growth, as measured in terms of growth percentiles, is required by each student to reach an official achievement target/goal established through the establishment of a future year’s achievement goal in the Star Growth Report, which allows for the specification of a desired proficiency level and what growth standard will be achieved to attain this objective. This approach can be applied to any number of future years or to a single year.
Unlike traditional mean/median analyses, SGPs are based on student-level performance and are therefore less susceptible to the same spurious correlations that plague other school level statistics. For example, when comparing school-level math SGPs from 2013, the mean and median SGPs varied less than the average difference between the school-level mean and the national median.
As a result, SGPs can provide more useful and meaningful information than traditional measures, such as averages. However, it is important to note that SGPs are only one tool in a toolbox of school improvement strategies and should be utilized alongside other measures.
While it is possible to perform some SGP analyses on WIDE formatted data, most higher level wrapper functions assume that the underlying SGP state data is stored in the LONG format and are therefore designed for use with this format. This is particularly true if you plan to run these analyses operationally, year after year, as the process of preparing and managing long data sets is much simpler than working with WIDE formatted data. For this reason, we recommend that you always work with LONG formatted data when performing SGP analyses.
